1

Author: Huang, Yanhui

Title: Applying Linear Programming Method in Six Sigma Approach to Develop a
Truck Planning Tool
The accompanying research report is submitted to the University of Wisconsin-Stout, Graduate School in partial
completion of the requirements for the
Graduate Degree/ Major: MS Technology Management

Research Adviser: James Keyes, Ph.D.

Submission Term/Year: December, 2013

Number of Pages: 56

Style Manual Used: American Psychological Association, 6th edition

I understand that this research report must be officially approved by the Graduate School and
that an electronic copy of the approved version will be made available through the University
Library website
I attest that the research report is my original work (that any copyrightable materials have been
used with the permission of the original authors), and as such, it is automatically protected by the
laws, rules, and regulations of the U.S. Copyright Office.
My research adviser has approved the content and quality of this paper.

STUDENT:

NAME Yanhui Huang DATE: December 16, 2013

ADVISER: (Committee Chair if MS Plan A or EdS Thesis or Field Project/Problem):

NAME Dr. James Keyes DATE:

---------------------------------------------------------------------------------------------------------------------------------
This section for MS Plan A Thesis or EdS Thesis/Field Project papers only
Committee members (other than your adviser who is listed in the section above)

1. CMTE MEMBER’S NAME: DATE:

2. CMTE MEMBER’S NAME: DATE:

3. CMTE MEMBER’S NAME: DATE:

---------------------------------------------------------------------------------------------------------------------------------
This section to be completed by the Graduate School
This final research report has been approved by the Graduate School.

Director, Office of Graduate Studies: DATE:

 2

Huang, Yanhui. Applying Linear Programming Method in Six Sigma Approach to Develop a

Truck Planning Tool

Abstract

Six Sigma is a tool for business to achieve continuous improvement. The benefits of

implementing Six Sigma projects include cost reduction and productivity improvement. Six

Sigma uses the Define, Measure, Analyze, Improve and Control (DMAIC) methodology to solve

problems. In this study, the linear programming approach was used in the improve and control

phases of a Six Sigma project to find a truck scheduling tool for ABC Transportation Company

to reduce their shipping cost. The key cost variables and their probability distributions were

identified in the analyze phase. The study results showed that the new truck scheduling tool

would reduce the shipping cost significantly.

 3

Acknowledgments

I would like to thank my husband, my son and other family members for their patience

and support during this project and throughout my studies. I would also like to thank ABC

Transportation Company for their support during this study. Finally, I would like to thank Dr.

Keyes and his assistant, Jamie Scholl, for their assistance and time.
 4

Table of Contents

............................................................................................................................................. Page

Abstract ...................................................................................................................................... 2

List of Tables .............................................................................................................................. 7

List of Figures ............................................................................................................................. 8

Chapter I: Introduction ................................................................................................................ 9

Statement of the Problem .................................................................................................... 10

Purpose of the Study............................................................................................................ 10

Definition of Terms ............................................................................................................. 11

Assumptions of the Study .................................................................................................... 11

Methodology ....................................................................................................................... 12

Limitations of the Study ...................................................................................................... 14

Summary ............................................................................................................................. 14

Chapter II: Literature Review .................................................................................................... 16

Six Sigma ............................................................................................................................ 16

DMAIC Methodology ......................................................................................................... 17

DMAIC Methodology and Continues Improvement ............................................................ 20

Linear Programming............................................................................................................ 21

The Relationship Between Chinese GDP Growth and Residential Building

Development ....................................................................................................................... 22

Truck Transportation Operation in China ............................................................................ 23

Summary ............................................................................................................................. 25
 5

Chapter III: Methodology .......................................................................................................... 27

Data Collection ................................................................................................................... 27

Data Verification ................................................................................................................. 28

Data Analysis – Pareto Analysis for High Cost Items .......................................................... 29

Data Analysis – Determine the Probability Distributions of Cost Items................................ 31

Development of Linear Programming Model....................................................................... 32

Solving the Linear programming Model .............................................................................. 34

Limitations .......................................................................................................................... 34

Summary ............................................................................................................................. 35

Chapter IV: Results ................................................................................................................... 36

Results of Data Verification ................................................................................................ 36

Results of Pareto Analysis ................................................................................................... 37

Results of Probability Distributions of Extra Cost Items ...................................................... 38

Linear Programming Model................................................................................................. 41

Results of Linear Programming Output ............................................................................... 42

Results of One Month Historical Data Comparison.............................................................. 44

Summary ............................................................................................................................. 44

Chapter V: Discussion............................................................................................................... 46

Limitations .......................................................................................................................... 47

Results ................................................................................................................................ 47

Conclusions ......................................................................................................................... 48

Recommendations ............................................................................................................... 49

Opportunities for Future Research ....................................................................................... 50

 6

References ................................................................................................................................ 53

Appendix A: ABC Transportation Company 2012 Shipping Cost Summary.............................. 56

 7

List of Tables

Table 1: Summary of Trucks Used for XYZ Electrical Business.................................................. 9

Table 2: Sample Truck Cost Record – Company Truck ............................................................. 27

Table 3: Sample Truck Cost Record – Rental Truck .................................................................. 28

Table 4: Sample Extra Cost Item Tracking Table – Company Truck ......................................... 30

Table 5: Sample Extra Cost Item Tracking Table – Rental Truck .............................................. 30

Table 6: Sample Overload Cost Probability Distribution - Overload vs. Overload Fine ............. 32

Table 7: Probability Distribution of Extra Costs for Rental Truck.............................................. 38

Table 8: Probability Distribution of Overload Weight vs. Overload Fine ................................... 39

Table 9: Probability Distribution of Overload Weight vs. Extra Fuel Cost ................................. 40

Table 10: Probability Distribution of Overload Weight vs. Late Delivery Penalty ..................... 41

Table 11: Actual Cost vs. Linear Programming Output Results for April 2012 Shipment .......... 44
 8

List of Figures

Figure 1: Cause and Effect Matrix Example .............................................................................. 19

Figure 2: DMAIC Six Sigma Improvement Methodology ......................................................... 20

Figure 3: Chinese GDP Annual Growth Rate from Year 2008 to Year 2012 .............................. 23

Figure 4: Chinese Diesel Prices Historical Data from Year 2005 to Year 2012 .......................... 24

Figure 5: Sample Size Code ...................................................................................................... 28

Figure 6: Sample Master Table .................................................................................................. 29

Figure 7: Example of a Pareto Chart for Cost Items Analysis .................................................... 31

Figure 8: Linear Programming Variable Constraint Set ............................................................. 33

Figure 9: Linear Programming Objective Function .................................................................... 34

Figure 10: Pareto Analysis Results for Key Cost Items .............................................................. 37

Figure 11: Linear Programming Model Variable Constraint Set ................................................ 42

Figure 12: Linear Programming Model Objective Function ....................................................... 42

Figure 13: MS Excel Solver Output – Linear Programming Model Solution .............................. 43

Figure 14: The Relationship between the Current Study and Future Research Opportunities ..... 50
 9

Chapter I: Introduction

A transportation service provider, hereafter referred to as ABC Transportation Company

(ABC) to protect the confidentiality of the information provided, which was established in 1997.

Based in Guangdong Province, China, the company provides transportation services throughout

mainland China. At the end of 2012, ABC had four departments and 52 employees. The main

customer of ABC Transportation Company was XYZ Electrical (an actual company, the name

has been altered to protect confidentiality).

The trucks used to ship the XYZ electrical products came from three resources; trucks

owned by ABC, trucks owned by drivers who were partnering with ABC and trucks rented by

ABC from other transportation service providers. The shipping costs of renting trucks from

other transportation service providers were significantly higher than that of ABC owned trucks

or drivers that are partnered with the company. If possible, ABC wanted to use its own trucks

and trucks owned by drivers that are partnered with the company. The rental option was only to

be used when the first two options could not meet the customers’ needs. Table 1 summarizes the

capacity, job assignment priority and operation cost of ABC’s truck resources.

Table 1

Summary of Trucks Used for XYZ Electrical Business

Owner Number Job Priority Operation Cost

Trucks owned by ABC 16 High Low

Trucks owned by sub-contractors 24 Medium Medium

Trucks rented Unlimited Low High

In the past, the majority of XYZ electrical products were shipped by trucks owned by

ABC and drivers partnered with ABC. Rental trucks were only used in emergency situations.
 10

With the dramatic development of real estate in China (new residential buildings increased on

average 7.7% per year from 2008 to 2012), the demand of XYZ Electrical products also

increased significantly (Ruan, 2012). As a result, XYZ Electrical increased the shipment of its’

products which included items such as: Wiring devices, home automation systems, structured

cabling systems and door entry systems. In order to meet the increased demand, XYZ Electrical

asked ABC to ship higher volume to more destinations and more frequently. ABC was not

prepared for the increase in demand and found it had to use more rental trucks from other

providers.

Statement of the Problem

ABC experienced a high number of late deliveries, unbalanced truckloads, and increased

use of rental trucks from other transportation service providers. This reduced the company's

profits.

Purpose of the Study

The purpose of the study was to use Six Sigma methodologies to improve ABC’s truck

capacity planning and scheduling to reduce the shipping cost and increase the company’s profits.

ABC used a demand driving scheduling process in which the truck scheduling was based on

customer demand. After receiving the customers’ order, the operations department combined the

orders and assignments into truckloads sorted by destinations. Trucks were assigned based on

the total number of truckloads required. If there were not enough company trucks available, ABC

rented trucks from contract truck providers. This approach came with the following issues; such

as, not enough trucks carrying products to specific destinations and lack of demand forecasting.

This study used a Six Sigma approach to reduce the unbalanced scheduling issues, so the

capacity of company trucks could be optimized.

 11

Definition of Terms

Six sigma methodology. Emphasize statistical tools to improve the established

processes (Reytblat, 2011).

DMAIC. Six Sigma problem solving steps: Define, Measure, Analyze, Improve, Control

(Hahn, & Hill, 1999).

Linear programming. The mathematic approach that is used to find optimized solutions

for the linear models with constraints (Moore & Weatherford, 2001).

Truckload. The total goods loaded to a single truck.

Overload. The load exceeds the designated truck capacity. This may cause delay of

delivery because the extra load can affect truck speed.

Target cost. The base cost of shipping one regular load truck to a certain destination.

The target cost is clean cost which means there are no overload fines and late delivery penalties.

Cost deviation. The difference between actual cost and target cost.

MIL-STD-105E. The acceptance sampling plan standard published by the Department

of Defense USA that emphasizes on verifying the accuracy of a population data set based on

sampling (Montgomery, 2009).

Pareto analysis. Used in identifying key problems (Montgomery, 2009).

Renminbi (RMB). Chinese currency, 1 RMB is equivalent to 0.16 US dollars.

Assumptions of the Study

Several assumptions were made throughout this study to find the optimized truck

schedule. First, ABC created a project team and provided sufficient resources to support the

project. The four team members represented company functions including; customer service,

scheduling, operation and accounting.

 12

Secondly, the operation data used within this study was accurate and up to date. Data

from 12 months operation was used to optimize the truck schedule. It was expected that the

historical data used by the project team was accurate. Sample operational data was selected

based on MIL-STD-105E to verify the accuracy.

Thirdly, the optimization model was created from the previous 12 months’ operational

data. This study assumed that the customer demand would remain at the same level; therefore,

the customer would continue working with ABC and there would be no significant change in the

shipment volumes. In this case, ABC did not have to purchase extra trucks to increase its’

resource capacity.

Finally, the geographic destinations of the shipments would remain in the same area.

Since the optimization scheduling model inputs were based on data from the previous 12 months

of operation; it was very important to keep the truck routes constant so the model could be used

for future scheduling.

Methodology

The objective of this study was to use Six Sigma methodologies to provide a planning

tool that resulted in lower costs of operation. It was accomplished by optimizing the truck

overload scheduling to lower the shipping cost. Linear programming was employed in the Six

Sigma analyze and improves phases in this study. Linear programming models contain an

objective function and several linear constraints. In this research, the objective function was to

minimize the cost in each operation cycle. The linear constraints included the total overload

shipment volume and the overload weight allowance of each truck. The total shipping cost was

calculated based on the probability distributions of the key cost items.

 13

The operational data was collected between January 2012 and December 2012. The data

included the following information:

1. Customers’ order data

The costumers’ order data contained the following information:

1) Total shipment volumes

2) Specific items for each order

3) Destination of the shipments

4) Desired delivery dates

2. Truck availability

The truck availability data included; how many trucks were available at each date, how many of

them were ABC owned trucks, how many of them were driven by partners of ABC, how many

trucks were under regular maintenance and how many trucks were not available due to

emergency issues such as accidents or driver illnesses.

3. Rental truck cost

The costs of rental trucks.

4. Actual shipment records

The actual shipment data included how each order was shipped, the delivery date and late

delivery penalties, overload fines, extra fuel costs and the actual costs of each order shipped.

The data were summarized by the following approach. First, the target cost was

determined for each destination area. Next, the deviations between the actual cost and target

costs were calculated. In the third step, the cost deviations were categorized into different cost

categories such as extra fuel costs caused by overload, overload fines, extra rental truck costs and

late delivery penalties.

 14

Pareto analysis was performed to examine the summarized data. The purpose of Pareto

analysis was to identify the main cost items that resulted in high operation costs. The linear

programming model was established based on the finding from the Pareto analysis. The objective

of the linear programming model was to minimize the shipping costs. Microsoft (MS) Excel

Server was used to solve the linear programming model.

The truck scheduling plan was formulated based on the results of MS Excel Server

outputs. The impact to the total shipping cost was calculated based on the recommended truck

schedule and was compared to the historical actual costs to see if there was a significant

improvement opportunity.

Limitations of the Study

This research project focused on utilizing the trucks owned currently by ABC

Transportation Company. Options such as increased resource capacity (buying new trucks or

adding more partner driver trucks) were excluded. Second, the data used in this research only

reflected the historical operational costs of ABC Transportation. Hence, the proposed specific

solutions from this research may not generalize to other businesses. Third, the forecasting model

used in this research did not apply to situations such as new competition entering the business or

government policy changes.

Summary

This chapter addressed the background of the research project including the problems

ABC faced, the purpose of the study and objectives of the research. It also defined important

terms and the limitations of the study.

The next chapter explores the literature related to the Six Sigma DMAIC approach, linear

programming and the transportation business in China. More importantly, the next chapter will
 15

review how to link the linear programming model to Six Sigma’s analyze and improve phases

and what benefits are associated with it.

 16

Chapter II: Literature Review

ABC Transportation Company experienced operational efficiency problems, which has

led to higher costs, late deliveries and poor truck utilization. The company wanted to utilize a Six

Sigma approach to reduce their operational costs and to increase their profitability. This

literature review analyzed Six Sigma methodology and the reason for using Six Sigma in this

project. It also reviewed linear programming and the benefit of applying linear programming to

improve ABC’s operational efficiency. In addition, literature related to the Chinese

transportation market was reviewed to provide an understanding about the research project.

Six Sigma

Six Sigma is “A business improvement approach that seeks to find and eliminate causes

of mistakes or defects in business processes by focusing on outputs that are of critical importance

to customers” (Snee, 1999, p. 100).

The American Society for Quality (2013) defines Six Sigma as:

a fact-based, data-driven philosophy of quality improvement that values defect

prevention over defect detection. It drives customer satisfaction and bottom-line results

by reducing variation and waste, thereby promoting a competitive advantage. It applies

anywhere variation and waste exist, and every employee should be involved.

Six Sigma is a well-defined process that focuses on problem solving which maximized the return

on a company’s investments. Six Sigma was developed by Motorola in the 1980’s to improve

manufacturing processes quality and to enhance customer satisfaction (He, Tang & Chang,

2010). At the Six Sigma quality level, with a 1.5 standard deviation shift from the process target,

the manufacturing process produces 3.4 defective parts per million opportunities (Harry, 1998).

Beginning in the 1990’s, organizations such as General Electric (GE) and AlliedSignal
 17

(Honeywell) used a Six Sigma approach to improve other aspects of operations such as reduced

cycle time, increased process yield, and enhanced resources efficiency (Evans & Lindsay, 2010).

The benefit of implementing Six Sigma projects include: Defect and cost reduction,

productivity improvements, cycle time reductions, customer relations improvements, and market

share increases (Pande, Roland & Cavanagh, 2000). Hahn & Hill (1999) reported that Motorola

saved $940 million over three years from Six Sigma projects. AlliedSignal (Honeywell) had an

estimated savings of $1.5 billion for the year 1997 from their Six Sigma initiative and GE

received $1.75 billion for 1998 and $2.5 billion for 1999 from its $1 billion investment in Six

Sigma improvement projects (Breyfoggle, 2003).

DMAIC Methodology

Six Sigma uses Define, Measure, Analyze, Improve and Control (DMAIC) methodology

to solve problems (Hahn & Hill, 1999).

The define phase is the first phase of the Six Sigma improvement process. In this phase,

the following tasks will be completed:

1. Define the stakeholders

The Six Sigma project Stakeholders included internal stakeholders and external stakeholders.

Internal stakeholders are within the organization such as top management team, functional

managers, and project team members. External stakeholders are outside the organization such as

customers, competitors, suppliers, and other intervener groups (Breyfoggle, 2003).

2. Analyze the impact of the Six Sigma project to the stakeholders

 18

The impact of the project to each stakeholder would then be analyzed. The project that is found

most beneficial to the organization would be selected. The tools used during the define phase are

Suppliers, Inputs, Processes, Outputs, and Customers (SIPOC) analysis (Breyfoggle, 2003).

3. Form the Six Sigma project team

After the project was selected, a Six Sigma project team was formed and roles were assigned to

each team member (Breyfoggle, 2003).

4. Develop Project Plan

The project team developed a project plan during the define phase. This project plan included

the problem statement, project goal, project scope, project team, project resources needed and

project timeline (Breyfoggle, 2003).

5. Approval

After the Six Sigma project plan was approved by the project sponsor, the resources would be in

place and the project team could move to the measure phase (Breyfoggle, 2003).

A process has to be measured before it can be improved. The measure phase is intended

to identify the data that illustrates the state of the current process. The Six Sigma team had to

first identify the input and output variables for the process. The team then developed a data

collection procedure and sampling plan (Breyfoggle, 2003). The data collection tools included

checklists, check sheets and a control chart (Evans & Lindsay, 2010). The Six Sigma team set a

baseline based on the data collected to determine the root causes for corrective actions.

Breyfoggle (2003), created a cause and effect matrix that served as a tool to help the Six Sigma

team prioritize the process input variables shown in Figure 1. The cause and effect matrix is used

through the following steps:

1. List the output variables with their prioritization number.

 19

2. Enter the effect value that each input variable has on the output variable.

3. The result values were calculated from the sum of the multiplication of effect value

and output priority.

4. The percentage impact of each input variable was calculated based on the

contribution of each input variable.

5. The results of the cause and effect matrix could help the Six Sigma team determine

which input variable affected the output variable the most.

Process Output Variables

A B C D Percentage
Prioritization number 4 2 3 1 Results Impact
1 3 1 3 3*4+1*3+3*1=18 18/71=25.4%
Process
2 5 7 5*2+7*3=31 31/71=43.6%
Input
Variables 3 2 4 2*4+4*1=12 12/71=16.9%
4 2 2 2*2+2*3=10 10/71=14.1%
Total 18+31+12+10=71 100%

Figure 1. Cause and Effect Matrix Example (Retrieved from Breyfoggle, 2003)

The analyze phase was the most important phase in the Six Sigma process (Eckes, 2003).

The improvement idea was generated based on the analyzed results. The recommendation for

improvement was made in this phase as well. The Six Sigma team used tools such as histograms,

Pareto analyses, experimental designs and statistical process control to find root causes for the

problem (Evans & Lindsay, 2010).

The improve phase is where the problem solution or corrective action was implemented.

The tools used in this phase included experimental designs and hypothesis testing (Evans &

Lindsay, 2010). This phase required multiple attempts until the project goals and objectives were
 20

reached. Once the team was able to show that the solution resulted in improvement, they could

move on to the control phase (Go Lean Six Sigma, 2013). Researchers recently used linear

programming as a mathematical and statistical tool in the Six Sigma analyze and improve phases

(Bobek, Imondi, Shott & Toobaei, 2012). The details about linear programming will be reviewed

later in this chapter.

The control phase is the last phase in the DMAIC procedure. The objective of this phase

is to sustain the achieved improvement in the improve phase. The Six Sigma team needed to

create a control plan to ensure the improved process maintained its ability to meet customers’

needs (Eckes, 2003). Statistical Process Control is often used in the control phase to monitor the

status of process.

DMAIC Methodology and Continues Improvement

Like Deming’s Plan-Do-Check-Act quality improvement cycle, Six Sigma DMAIC

procedure is a tool for business to achieve continued improvement (Evans & Lindsay, 2010). An

organization starts a new Six Sigma improvement after one project is completed. Figure 2 shows

the DMIAC methodology. Plan-Do-Check-Act is considered the foundation of DMAIC. There

are many similarities between Plan-Do-Check-Act and DMAIC.

Figure 2. DMAIC Six Sigma Improvement Methodology

 21

Linear Programming

Linear programming is the process of finding the optimized outcome for a given set of

linear equations related to a specific situation (Moore & Weatherford, 2001). Many real world

problems lend themselves to linear programming modeling or can be approximated by linear

models. In general, a linear model contains three types of functions:

1. Linear objective function

This objective function defined the goal of the linear model. Each linear model can only have

one goal such as maximum profit or minimum cost, shortest processing time and largest amount

of products throughout (Moore & Weatherford, 2001).

2. Linear constraint set

The linear constraint set defined the situational constraints such as; number of resources

available, total amount of time an order has to be completed and machine hours between regular

maintenance. For linear programming models, all constraints have to be in linear functions

(Moore & Weatherford, 2001).

3. Variable constraint set

The variable constraint set indicated the variable range. Each variable must has its own

constraint defined in the constraint set (Moore & Weatherford, 2001).

Linear programming can be solved by using computer software such as MS Excel Solver.

However, not all the linear programming models can solve for optimized solutions under the

following three situations:

1. Infeasibility

The linear constraint set is unsolvable, which means there is no variable combination that meets

all constraints (Moore & Weatherford, 2001).

 22

2. Unsoundness

The objective function had a result close to infinity (Moore & Weatherford, 2001).

3. Alternate solution

There was more than one variable combination that resulted in the same object function result

(Moore & Weatherford, 2001). In order to avoid the above situation, researchers often modified

the linear model to reach a unique optimized solution. Industry has had a long history of

applying linear programming to solve complex problems and find the optimized solutions in

production operation and services areas (Balbirer & Shaw, 1981). Linear programming was used

to find optimized solutions for complex situations; researchers recently started to use it as a

mathematical and statistical tool in the Six Sigma analyze and improve phases (Bobek et al.,

2012).

The Relationship Between Chinese GDP Growth and Residential Building Development

According to Dickinson (2012), GDP stands for Gross Domestic Product, which is the

market value of all goods and services produced within a country in a given period of time. From

1989 to 2012, the Chinese average GDP Annual Growth Rate was 9.2% (National Bureau of

Statistics of China, 2013a). Figure 3 shows the GDP annual growth rate from 2008 to 2012 in

China.

Investment in new residential building development contributed heavily to Chinese GDP

growth (Sahoo, Dash & Nataraj, 2010). The dramatic growth of new residential development

generated increased demand of raw materials, as well as the transportation of these materials.

ABC is one of the beneficiaries of this demand increase.

 23

Figure 3. Chinese GDP Annual Growth Rate from 2008 to 2012 (National Bureau of Statistics of

China, 2013b)

Truck Transportation Operation in China

In general, there are two types of truck transportation service providers in the Chinese

market. They are government owned providers and private providers. ABC would be considered

as a private provider. Unlike the government owned providers, which have business from

government owned companies; ABC needs to sell its transportation service to customers and

produce a profit to stay in business.

At the beginning of each year, ABC and its’ customers negotiate a contract for the

transportation service. The key element of the contract is the unit price, which is the price that

customers pay for the shipment of a certain amount of goods to a certain destination within a

certain amount of time. Once this unit price is agreed upon by both parties; it normally would not

change until the next year unless one party decides to stop the business relationship. In addition,

the competition in the transportation market was getting stronger. There were 20% more private

truck transportation service providers in 2012 than 2011 in Guangdong Province (Guangdong
 24

Bureau of Statistics of China, 2013). The operational costs of a private truck transportation

provider include the following five categories: fuel costs, toll costs, overload fines, late delivery

penalties and truck rental costs.

The first category is fuel costs, the diesel prices increased 80% from 2009 to 2012; Figure

4 shows the increase of fuel prices in China (National Development and Reform Commission,

2012). The second item is the toll costs. Vehicles have to pay toll costs while using these

highways because the highways were built by funding from banks. Chinese banks belong to the

Government, which defines the toll price. The Government will typically develop a long term

toll price to recover the invested costs of the highway system (Chen, 2012).

Figure 4. Chinese Diesel Prices Historical Data from 2005 to 2012 (National Development and

Reform Commission, 2012)

For instance, a truck would have to pay RMB 1,180 or $180 toll fee from Huizhou, Guangdong

Province to Wuxi, Zhejiang Province (around 1,400 kilometers).

 25

The third cost item is the fine of overload. In order to offset the increasing fuel and

highway costs, many transportation providers reconfigure their trucks so they can overload

goods shipped on their trucks. A study showed that 100% of the Chinese trucks in Southern

China had been overloaded. Some trucks had overloaded the normal capacity, by as much as

800% (Chen, 2012). Most of the time, the overloaded trucks just pay a fine and continue to their

destination (Wei, 2012).

The fourth cost item is the late delivery penalties. Large amounts of overloaded trucks

slowed down the traffic and increased the probability of accidents, which resulted in late

deliveries. Customers charged late delivery penalties to the trucking firm. The transportation

company then needed to absorb the cost of the penalty.

The fifth cost item is renting from other transportation providers. In general,

transportation providers charge higher prices than company owned trucking costs. This price

may have even been higher than what customers were willing to pay to ABC.

Summary

This chapter covered the Six Sigma process improvement approach, the linear

programming method and the Chinese truck transportation business. The Six Sigma approach is

often used to reduce costs and increase operational efficiency. Linear programming can help the

organization determine the best operational variable combination, which leads to improved

profit. Given the operational challenge ABC faced, this research intended to use the above

approaches to help the company find a way to better utilize its resources and reduce their costs.

The next chapter will look at the methodology associated with Six Sigma process

improvement and linear programming resource optimization. It will go over the methods
 26

employed to help ABC find their best operational strategy. The data collection and procedures

used in the research will also be covered.

 27

Chapter III: Methodology

ABC Transportation Company experienced a high number of late deliveries, unbalanced

truckloads, and increased use of rental trucks from other transportation service providers. This

reduced the company's profits. The objective of this study was to use Six Sigma methodologies

to provide a planning tool that resulted in lower cost of operation. Linear programming was used

in the Six Sigma analyze and improved phases in this study. The linear programming model

contained an objective function and several linear constraints. In this research, the objective

function was to minimize the cost deviation in each operation cycle. The linear constraints

included the total overload weight and the truck overload capacity.

Data Collection

The shipping cost records were collected for 2012. The cost records included shipping

destination, date shipment requirement received, delivery date, total overload weight and total

cost. The cost records were entered into an MS Excel file. An identification number was

assigned to each record. There were two types of shipping cost records. The first type of

shipping cost records were the shipping costs for the trucks owned by ABC and their sub-

contractors. Table 2 shows an example of this type of record.

Table 2

Sample Truck Cost Record – Company Truck

ID 10
Destination Wuxi City
Target Cost RMB 5,000 or $840
Date Order Received 01/15/2012
Delivery Date 01/30/2012
Total Overload Weight 7 Ton
Total Cost RMB 6,400 or $1,060
Rental Truck? No
 28

The second type of shipping cost records were the shipping costs for the rental trucks. An

example of rental truck shipping costs is shown in Table 3.

Table 3

Sample Truck Cost Record – Rental Truck

ID 12
Destination Wuxi City
Target Cost RMB 5,000 or $840
Date Order Received 01/15/2012
Delivery Date 01/31/2012
Total Overload Weight 10 Ton
Total Cost RMB 8,200 or $1,360
Rental Truck? Yes

Data Verification

The accuracy of the truck cost records were checked before this data was analyzed. MIL-

STD-105E sampling plan defined by the U.S. Department of Defense was used to determine the

sample size (Montgomery, 2009). In the first step, Sample Size Code Letter was determined

from the MIL-STD-105E standard shown in Figure 5.

Figure 5. Sample Size Code (Montgomery, 2009)

 29

Secondly, the total number of samples that needed to be checked was identified from the

MIL-STD-105E Master Table shown in Figure 6. The acceptance level was also determined

from the Master Table. In the third step, the researcher used the MS Excel random number

generator to determine which record would be selected for verification. Finally, the selected

sample records were checked to verify if the truck cost record data met the accuracy standard. If

the verification results showed that the data accuracy met the MIL-STD-105E standard, the data

set could be used for analysis. If the verification results show the data accuracy met the MIL-

STD-105E standard, a 100% double check of the shipping cost records had to be done to ensure

all the errors in the data set were corrected before any data analysis was conducted.

Figure 6. Sample Master Table (Montgomery, 2009)

Data Analysis – Pareto Analysis for High Cost Items

Pareto analysis was used to analyze data to determine the key cost items that affected

truck cost. ABC established target cost for each destination. The target cost was the base cost of
 30

shipping one regularly loaded truck. The target cost was a clean cost, which means there were no

overload fines, no extra fuel costs and no late delivery penalties. Table 4 shows an example of

how the extra cost items were collected for the company owned trucks.

Table 4

Sample Extra Cost Item Tracking Table – Company Truck

ID 10 Extra Cost
Destination Wuxi City
Target Cost RMB 5,000 or $840
Date Order Received 01/15/2012
Delivery Date 01/30/2012
Total overload 7 Ton
Total Cost RMB 6,400 or $1,060
Rental Truck? No
Extra Cost from Rental 0 0
Extra Fuel Cost RMB 620 or $100 RMB 620 or $100
Extra Toll Cost 0 0
Overload Fine RMB 660 or $110 RMB 660 or $110
Late Delivery Penalty RMB 100 or $17 RMB 100 or $17
Other Extra Cost RMB 20 or $3

Table 5 shows how the extra cost items were collected for rental trucks. For rental

trucks, the rental company had to pay extra fuel costs, extra toll costs, overload fines and late

delivery penalties.

Table 5

Sample Extra Cost Item Tracking Table – Rental Truck

ID 12 Extra Cost
Destination Wuxi City
Target Cost RMB 5,000 or $840
Date Order Received 01/15/2012
Delivery Date 01/31/2012
Total Overload Weight 10 Ton
Total Cost RMB 8,200 or $1,360
Rental Truck? Yes
Extra Rental Cost RMB 3,200 or $520
 31

A Pareto chart was developed to determine the top contributing extra cost items. Figure 7

shows an example of a Pareto chart which ranked extra cost items from the highest to lowest

contribution. The actual results will be shown in Chapter IV. The key contribution cost items

were used as decision variables for the linear programming model.

Figure 7. Example of a Pareto Chart for Cost Items Analysis

Data Analysis – Determine the Probability Distributions of Cost Items

As mentioned in Chapter II, most Chinese transportation service providers overloaded the

weight capacity of their trucks to reduce shipping cost. The consequences associated with

overloaded trucks were extra costs including; possible overload fines, extra fuel costs and

potential late delivery penalties. However, since there were different random factors such as

number of traffic control agencies on duty that might affect the values of extra cost items (e.g. a

truck overloaded by five tons to Wuxi City had to pay an overload fine of RMB 660 or $110 for

one trip and RMB 850 or $140 for another), it was necessary to understand the probability

distribution of the extra cost items. In this study, the researcher determined the probability

distribution of the extra cost items based on the frequency of occurrences in the truck record

data. Table 6 shows an example of the cost probability distributions between overload and
 32

overload fine. The probability distributions for other key cost items such as the probability

distributions between overload and late delivery penalty, overload and extra fuel cost will also be

determined using the same methodology. The results of probability distributions for the key cost

items will be shown in Chapter IV.

Table 6

Sample Overload Cost Probability Distribution - Overload vs. Overload Fine

Overload (Ton) Overload Fine Probability

Under 5 RMB 500 or $80 50%

RMB 740 or $120 35%
RMB 800 or $130 15%
5-8 RMB 500 or $80 10%
RMB 740 or $120 50%
RMB 800 or $130 40%
8-10 RMB 740 or $120 15%
RMB 800 or $130 45%
RMB 1,000 or $160 40%

Development of the Linear Programming Model

Linear programming was used during the improve phase of the DMAIC process to

determine the truck scheduling policy. As mentioned in Chapter II, linear programming has a

history of being implemented as a tool to help business sectors determine the optimized

utilization of their resources. The linear programming model was built through the following

steps:

1. Variables
 33

The variables in this model were the overload weights of truckloads for each shipment. These

variables were defined as V1 to Vn, which stands for the total overload weight of each truck.

2. Variable Constraint Set

The first set of variable constraints was the overload weights of each truck; the overload weights

should not contain negative numbers. The second constraint was the total overload weight

volume. This volume will be greater than or equal to the total shipment overload volume.

The third constraint was the overload constraint. Even though it was common in the Chinese

transportation market that trucks were overloaded, each transportation service provider

established a maximum overload weight allowance that a truck could be overloaded for safety

concerns. ABC had its’ own limit for how much weight a truck could handle for different

destinations. Based on the above constraint regulations, the following constraint set was

established. The constraint set is shown in Figure 8.

Figure 8. Linear Programming Variable Constraint Set

3. Linear Objective Function

The objective of this research project was to reduce the shipping cost for ABC Transportation.

Based on this, the objective function for linear model was established to minimize the total cost

as shown in Figure 9.
 34

Figure 9. Linear Programming Objective Function

Solving the Linear programming Model

The linear programming model was developed using MS Excel Solver, which would

solve the model and return the best combination of values for V1 through Vn. The results will

be shown in Chapter IV.

The project team used one month’s historical data to validate if this approach would

reduce the shipping costs. The truck costs developed through the linear programming model was

compared with the previous year’s costs to determine if there was a need to modify the model.

Limitations

As pointed out in Chapter I, there were several limitations for this study. First, the

project team only analyzed 12 months of operational data for this analysis, which implied that

the economic growth impact might not have been addressed in the linear programming.

The second limitation of this study was the probability distributions of extra cost items.

They were based on past experiences. With changes of road conditions such as new highways,

the parameters in the linear programming models might have to be modified to reflect the current

conditions. While the project provided a framework for optimized truck scheduling, it was

limited to the costs that were current at the time of the research.

The same limitation applied to the probability distribution of overload fines. While the

team tried to calculate the probability distributions of overload fines, the accuracy of these
 35

probabilities was limited to the information provided by the truck cost records during the

research study period.

The fourth limitation of this study was the lack of forecasting for future demand. Given

that residential building development was growing at a fast pace in China, there is a higher

probability that the demand from customers may increase in the near future. This trend was not

addressed in the linear programming model.

Summary

This chapter covered the research methodology for evaluating the shipping costs of ABC

Transportation. Pareto analysis was used to determine the key extra cost items. The probability

distributions of extra cost items were also identified.

Next, this chapter covered how to use the parameters generated from the data analysis to

establish the linear programming model. The objective function of the linear programming

model was to minimize the total extra cost. The parameters were used in the objective function

as well as the linear programming constraints. Verification was done based on one month of

operational data. A new policy was established to ensure the improved performance was

sustained.

Finally, the limitations of the study were given. The main concern of this study was the

accuracy of the probability distribution for the extra cost items. These distributions may change

due to the fast development of the Chinese economy. The linear programming model parameters

would have to be updated if these changes become significant.

The next chapter will review the results of this study. The linear programming model

outcome will be discussed in detail in the next chapter as well. Truck scheduling policy related to

the linear programming model output will also be explained.

 36

Chapter IV: Results

ABC Transportation Company experienced a high number of late deliveries, unbalanced

truckloads, and increased use of rental trucks from other transportation service providers. This

has increased the company’s shipping costs and reduced their profits. The goal of this study was

to use Six Sigma methodologies to provide a truck capacity and schedule planning tool that

would help ABC lower its shipping cost.

To accomplish the goal, linear programming was used in the Six Sigma analysis and

improves phases in this study to develop a truck capacity and scheduling plan. The research was

conducted through the following phases. First, the shipping cost records for 2012 were collected

and verified for accuracy. Pareto analysis was conducted in the second step to determine the key

cost variables that affected the shipping costs. Third, the probability distributions of the key cost

variables (extra rental cost, overload fines, late delivery penalties and extra fuel costs) were

determined. In the fourth step, key variables found from the second step and their probability

distributions identified in the third step were used to build the linear programming model.

Lastly, the results of linear programming model from MS Excel Solver were used to improve

ABC’s truck capacity planning.

Results of Data Verification

There were total of 816 truck cost records collected for this study. According to MIL-

STD-105E standard Sample Size Code Letter table showed in Chapter III Figure 5, the sampling

plan letter code was letter K. The total number of samples needs to be checked with letter code

K was 125 based on MIL-STD-105E Master Table showed in Figure 6 in Chapter III. The master

table also indicated that the acceptance level of this verification should be 0 which meant if there

was no error in 125 samples, the data set would be considered as accurate. The researcher
 37

selected 125 shipping records from the 816 data set and verified the accuracy of them. The result

showed that all of them were correct. The data set was considered accurate and could be used

for analysis.

Results of Pareto Analysis

Pareto analysis was used to identify the key cost items that affected the total truck cost.

These variables were programmed into the linear programming model.

The input cost items for the Pareto analysis were, extra fuel cost caused by overloaded

trucks, overload fines, late delivery penalties, extra cost of rental trucks and extra toll costs

caused by overload. The cost items of 816 shipping records were summarized (summary results

shown in Appendix A). The Pareto chart shown in Figure 10 was created based on the data

summary results.

Figure 10. Pareto Analysis Results for Key Cost Items

The results of Pareto analysis showed that the extra toll costs only contributed 1.1% to

the total cost, which meant it would not be considered as a key variable when making the truck

planning tool. The other four cost items (extra rental, overload fine, extra fuel cost and late
 38

delivery penalty) made significant contributions to the total cost and were included in the linear

programming model.

Results of Probability Distributions of Extra Cost Items

The Pareto analysis indicated that there were four cost items that significantly affected

ABC’s shipping cost. These four cost items were extra cost of rental truck, extra fuel cost,

overload fine and late delivery penalty. The probability distributions of each cost item were

needed to build the linear programming model. In this research project, the probability

distributions were determined by frequency count.

ABC Transportation Company incurred extra costs to use rental trucks. The extra costs

rental companies charged varied depending on the market. According to the frequency count

result, ABC used 280 rental trucks in 2012. ABC paid RMB 4,500 or $725 for 168 of them,

RMB 4,800 or $770 for 98 of them and RMB 5,200 or $840 for the rest 14 of the rental trucks.

Based on this frequency count outcome, the probability distribution of extra charge per truck was

calculated and summarized in Table 7.

Table 7.

Probability Distribution of Extra Costs for Rental Truck

Extra Charge by Rental Probability

Company (per Truck)
RMB 4,500 or $725 168/280 = 60%
RMB 4,800 or $770 98/280 = 35%
RMB 5,200 or $840 14/280 = 5%

The Pareto analysis results shown in Figure 8 indicated that there were three key cost

items associated with using overloaded company owned trucks: Extra fuel costs, overload fines

and late delivery penalties.

 39

A frequency count approach was used to determine the probability distribution between

overload weights and overload fines. For example, the 2012 shipping cost records showed that

there were 75 trucks that had overload weights between two to three tons. Among these 75

trucks, 15 trucks (20%) were charged with RMB 300 or $50 overload fine, 30 trucks (40%) were

charged with RMB 400 or $65 overload fine and the remaining 30 trucks (40%) were charged

with RMB 600 or $80 overload fine. This probability distribution was used in the linear

programming model to calculate the overload fines of trucks that had been overloaded by

between two to three tons. Table 8 shows the probability distribution between overload weight

(ton) and overload fines.

Table 8

Probability Distribution of Overload Weight vs. Overload Fines

Overload Weight (Ton) Overload Fine Probability

1 RMB 200 or $35 100%
2-3 RMB 300 or $50 20%
RMB 400 or $65 40%
RMB 600 or $80 40%
4-5 RMB 600 or $80 20%
RMB 800 or $130 45%
RMB 1,000 or $160 20%
RMB 1,200 or $190 15%
6-7 RMB 800 or $130 15%
RMB 1,200 or $190 50%
RMB 1,400 or $225 25%
RMB 1,600 or $260 10%
 40

The same approach was used to determine the probability distribution between overload

weight and extra fuel costs. Table 9 shows the probability distribution between overload weight

and extra fuel costs.

Table 9
Probability Distribution of Overload Weight vs. Extra Fuel Cost
Overload Weight (Ton) Extra Fuel Cost Probability
1 RMB 850 or $140 45%
RMB 900 or $145 55%
2 RMB 900 or $145 60%
RMB 1,000 or $160 40%
3 RMB 1,100 or $170 30%
RMB 1,150 or $185 40%
RMB 1,200 or $193 30%
4 RMB 1,150 or $185 40%
RMB 1,250 or $200 50%
RMB 1,350 or $220 10%
5 RMB 1,250 or $200 15%
RMB 1,350 or $220 20%
RMB 1,400 or $225 30%
RMB 1,450 or $235 35%
6 RMB 1,400 or $225 35%
RMB 1,450 or $235 30%
RMB 1,550 or $250 35%
7 RMB 1,450 or $235 20%
RMB 1,550 or $250 35%
RMB 1,650 or $265 35%
RMB 1,700 or $275 10%
 41

Table 10 shows the probability distribution between overload weight and late delivery

penalties. The probabilities were calculated from frequency count results.

Table 10

Probability Distribution of Overload Weight vs. Late Delivery Penalty

Overload Weight (Ton) Late Delivery Penalty Probability

2-3 RMB 200 or $35 15%
RMB 400 or $65 65%
RMB 600 or $100 20%
4-5 RMB 200 or $35 10%
RMB 400 or $65 25%
RMB 600 or $100 55%
RMB 800 or $130 20%
6-7 RMB 400 or $65 10%
RMB 600 or $80 10%
RMB 800 or $130 45%
RMB 1,000 or $160 35%

Linear Programming Model

The key cost items and their probability distributions were used to build the linear

programming model. The company’s maximum overload weight allowances were also

determined. For ABC’s company owned trucks, the maximum overload weight allowance is

seven tons. The maximum overload weight allowance for rental trucks is 10 tons. The linear

constraint set of the linear programming model is shown in Figure 11.

 42

Figure 11. Linear Programming Model Variable Constraint Set

The goal of this research was to find a scheduling tool to reduce the shipping cost. In

order to achieve this goal, the linear objective function was used to minimize the total shipping

cost as shown in figure 12.

Figure 12. Linear Programming Model Objective Function

Results of Linear Programming Output

MS Excel Solver was used to solve the linear programming model. The objective

function of the linear model was to minimize the total extra costs. The total extra costs were the

summary of extra fuel costs, overload fines, late delivery penalties and extra rental costs. The

output of MS Excel solver provided the specified overload value (ton) of each truck. Figure 13 is

a screenshot of a MS Excel Solver output.

For this particular week, ABC had to ship 108 tons of overloaded products. The linear

programming solution made the following recommendations:

 43

1. The output showed that company trucks V1 through company trucks V10 had final

values greater than zero. This indicated that the company had to use 10 company

trucks.

2. The overload value of each company truck was the final value from the output.

3. The output showed that the final values of rental trucks V1 through V5 were greater

than zero, which indicated that ABC had to use five rental trucks.

4. The overload value of each rental truck was the final value from the output.

5. The total overload amount was 108 tons, which equals the customer requirement.

Figure 13. MS Excel Solver Output – Linear Programming Model Solution

 44

6. The total cost of the week was RMB 58,450 or $9,430 based on this truck capacity

planning.

Results of One Month Historical Data Comparison

A pilot test was conducted by using four weeks of historical data from April 2012 with

the linear programming approach. The results in Table 11 show that the total truck costs under

this new planning approach would have been RMB 240,739 or $38,828, which was 15.2% less

than the actual truck cost (RMB 284,025 or $45,810) of that month. The new approach reduced

the truck cost significantly.

Table 11

Actual Cost vs. Linear Programming Output Results for April 2012 Shipment

Cost Item Actual Cost Cost Based on Linear Programming

(April 2012)
Extra Rental Cost RMB 99,400 or $16,030
Extra Fuel Cost RMB 71,045 or $11,460
Late Delivery Penalty RMB 62,480 or $10,080
Overload Fine RMB 51,100 or $8,240
Total RMB 284,025 or $45,810
Output (April, 201)
RMB 84,200 or $13,580
RMB 60,239 or $9,720
RMB 53,100 or $8,560
RMB 43,200 or $6,970
RMB 240,739 or $38,828

Summary

This chapter began by showing the results of truck cost data verification. It showed that

the data used in this research was accurate. The chapter also went over the results of Pareto

analysis, which was used to identify the key cost items. These items were used as decision

variables to build the linear programming model. After the key cost items were decided, the

probability distributions between the overload and the key cost items were identified. The

probability distributions were used in the linear programming model for truck cost calculations.
 45

MS Excel Solver was used to solve the linear programming model. The company’s operation

team used the linear programming model solution to lead the truck capacity planning. The pilot

test results showed that this approach would reduce the shipping cost for ABC Transportation

Company. The next chapter will draw conclusions from the study and make recommendations.

The opportunities for future research will also be discussed in the next chapter.
 46

Chapter V: Discussion

Six Sigma is a tool used to achieve continued improvements. The benefits of

implementing Six Sigma projects include; cost reduction and productivity improvement. Six

Sigma uses Define, Measure, Analyze, Improve and Control (DMAIC) methodology to solve

problems. In this study, a linear programming approach was used during the improve and

control phases of a Six Sigma project to develop a truck scheduling tool for ABC Transportation

Company.

Chapter I introduced the background of ABC Transportation Company and the

challenges of truck scheduling they were facing. ABC was unable to effectively plan for truck

scheduling which resulted in high shipping costs. The goal of this study was to use linear

programming methodology and Six Sigma approach to help ABC Transportation Company

create a truck scheduling tool. This chapter also covered the assumptions of the study.

Chapter II was a literature review of Six Sigma approach, linear programming

methodology, Chinese economic development and the truck transportation service in China. The

chapter covered Six Sigma DMAIC and why it could help an organization reduce costs and

improve operational efficiency. It also described linear programming in depth and how an

organization may benefit from using such a technology.

Chapter III covered the methodology for this study regarding collection and analysis of

historical data, built a linear programming model and provided truck scheduling tools for ABC

Transportation. The 2012 truck cost records were collected for this study. These records were

verified by MIT-STD-105E for accuracy. Pareto analysis was then performed to determine the

key cost variables that affected the truck shipping costs. The probability distributions of the key

cost variables were identified based on frequency count. A linear programming model was built
 47

with the key cost variables and their probability distributions. The objective of the linear

programming model was to minimize the total shipping cost. Chapter III concluded with the

discussion of the limitations of the study.

In Chapter IV, the results of data verification, Pareto analysis, probability distributions of

key cost variables and the linear programming output were presented. The linear programming

output was used as ABC’s truck scheduling tool. The one month pilot test result was conducted

and the shipping costs from the linear programming output were compared with the actual cost.

The comparison results showed that the cost based on the proposed scheduling tool was

significantly lower than the actual cost, which meant the new scheduling tool could reduce

ABC’s shipping cost.

This chapter draws conclusions from this research project followed by recommendations

that are made for ABC Transportation Company. Opportunities for future research will be

discussed at the end.

Limitations

This study was limited to the ABC Transportation Company’s 2012 operation records so

the economic growth impact factor could not be identified. The second limitation of this research

was the probability distributions of the key cost variables were also determined only by 2012

operation data which means the changes of road conditions, fuel price and administration

policies would not be reflected.

Results

There were 816 shipping cost records in 2012. The data verification results showed that

these records were accurate and could be used in this research project.
 48

The results of Pareto analysis showed that there were four cost items that had significant

contributions to the total shipping cost. These four key cost items included extra rental cost, extra

fuel cost, overload fine and later delivery penalties. The probability distributions for these key

cost items were identified based on frequency count.

The linear programming objective function took into account the above key cost items

and probability distributions and summed them for the total shipping cost. The goal of this

research was to find a scheduling tool to reduce the shipping cost. In order to achieve the

research goal, the linear objective function was set up to minimize the total shipping cost. The

linear constraint functions in this research include total overload shipment weight and maximum

overload allowance of each truck. The variables in the linear programming model were the

overload weights of each truck.

The linear programming model was solved by MS Excel Solver. In order to test if the

outcome of the linear programming model could help ABC Transportation Company reduce its

shipping cost, a one month pilot test was conducted. The shipping costs were calculated from

linear programming output and were compared with the actual shipping costs for that month.

The comparison showed that the linear programming model reduced the shipping cost

significantly.

The pilot test results indicated that the linear programming model could help the ABC

Transportation better schedule its truck capacity to reduce shipping cost. This would increase

the company’s profit level.

Conclusions

The literature review demonstrated that Six Sigma DMAIC approach could reduce an

organizations operational cost. It showed that organizations can benefit from successful Six
 49

Sigma projects. Linear programming can help an organization determine the optimized capacity

planning and resource scheduling. The literature review also showed the challenges and

opportunities that Chinese transportation providers are facing.

A goal of this study was to focus on providing a truck scheduling tool that ABC

Transportation Company could use to reduce the shipping cost. This goal was accomplished

through Six Sigma DMAIC approach. The key cost variables and their probability distributions

were identified in the analyze phase. Linear programming was used in the improvement phase to

minimize the total cost.

The findings in this study correlate to the literature review. For ABC transportation

Company, there would be a significant cost savings with the use of linear programming

scheduling tool.

Recommendations

Based from the results of this study, the following recommendations are being made for

ABC Transportation Company. First, the company should start using the new scheduling tool

from this research project. This new tool will help the company reduce the shipping cost.

Next, it is being recommended that ABC Transportation Company overload the trucks from

rental companies to their full overload weight limit. The linear programming output showed that

the overload weights for rental trucks equaled their overload limit. It is believed that this

strategy will reduce the risks of high overload fines, high late delivery penalties and high fuel

costs to the rental company. As part of this study, it is also recommended that ABC needs to

negotiate with the rental companies for a fixed price rental agreement to avoid rental cost

increases. Third, ABC should keep updating the linear programming model with the most current
 50

shipping cost records. It is recommended that ABC update the probability distributions of key

cost variables so the linear programming model would reflect the current road situation.

Finally, a future recommendation would be for ABC transportation Company to

continually apply the Six Sigma approach to search for additional opportunities to reduce the

shipping cost and improve their operation effectiveness.

Opportunities for Future Research

In the research project, ABC Transportation Company used linear programming model as

a truck scheduling tool to determine the overloaded weight of each truck. The results of the pilot

test showed that this could reduce the shipping cost for ABC significantly. There are future

research opportunities that linear programming model could help ABC improve its operation.

Figure 14 shows the future research project opportunities that ABC Transportation

Company may use to improve its operation.

First (current) Level Second Level Third Level

Linear Programming Model Linear Programming Model Linear Programming Model
Focusing on truck overload Adding other internal cost Adding external cost variables to
scheduling to reduce ABC’s variables to reduce ABC’s help ABC establish its business
shipping cost. the total operation cost. strategic plan.

Figure 14. The Relationship between the Current Study and Future Research Opportunities

First, the linear programming model in this study was built based on two sources, the key

cost variables from the Pareto analysis and the probability distributions of these key variables.
 51

An additional study focusing on the key variable probability distributions may be needed. The

higher the accuracy of the probability distributions, the closer the linear programming model

reflects the actual operational situation. This will increase the confidence level of using the

linear programming outputs.

Secondly, this project only focused on using linear programming model as a truck

scheduling and capacity planning tool to determine the overload weight of each truck. It might

be worthwhile in the future research projects to evaluate ABC Transportation Company’s total

cost, which include other cost items such as purchasing trucks for the company, maintenance

cost for company owned trucks and salaries and benefits of company owned truck drivers. This

would allow the linear programming model to represent the company’s operational costs more

accurately.

Thirdly, a future study might be conducted to evaluate ABC Transportation Company’s

business model. The literature review indicated that the competition of the Chinese truck

transportation service market is getting stronger. ABC should perform a strengths, weakness,

opportunities and threats (SWOT) analysis and reevaluate their business model to determine if

they need to change their strategic plan to become a marketing company and outsource the truck

shipping services. A linear programming model could be used in this type of evaluation. The

external factors that impact the operational costs could be used in building the linear

programming model. The output of the linear programming model would help the management

team figure out what business model will bring the most financial benefit to the company.

Another research opportunity could be the deeper evaluation of the linear programming

outputs. Future researchers may use linear programming sensitivity analysis to evaluate different

what-if truck scheduling situations in the linear programming outputs. However, in order to
 52

perform this type of evaluations, more powerful linear programming software such as LINGO®

will be needed for this type of analysis.

 53

References

American Society for Quality. (2013). Six Sigma learning resources & information. Retrieved

from http://www.asq.org/learn-about-quality/six-sigma/overview/overview.html

Balbirer, S., & Shaw, D. (1981). Application of linear programming to bank financial planning.

Journal on the Practice of Operations Research, 11(5), 77-83.

Bobek, A., Imondi, C., Shott, T., & Toobaei, M. (2012). Heterogeneous project scheduling for

optimal Six-Sigma cost reduction using linear programing. Technology Management for

Emerging Technologies (PICMET), 2012 Proceedings of PICMET '12, 35-42. Retrieved

from http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=6304257&url=http %3A

%2F%2Fieeexplore.ieee.org%2Fiel5%2F6296788%2F6304012%2F06304257.pdf%3Far

number%3D6304257

Breyfoggle III, F. W. (2003). Implementing Six Sigma – Smarter solutions using statistical

methods. Hoboken, NJ: John Wiley & Sons.

Chen, Z (2012). Inspection result shows 100% trucks were overloaded. Retrieved from

http://news.dayoo.com/guangzhou/201008/12/73437_13564655.htm

Dickinson, E. (2012). GDP: A brief history. Retrieved from http://www.foreignpolicy.com/

articles/2011/01/02/gdp_a_brief_history#sthash.1tF4JHfh.dpbs

Eckes, G. (2003). Six Sigma for everyone. Hoboken, NJ: John Wiley & Sons, Inc.

Evans, J., & Lindsay, W. (2010). Managing for quality and performance excellence.

Independence, KY: South-Western Cengage Learning.

Go Lean Six Sigma. (2013). DMAIC – The 5 phases of lean Six Sigma. Retrieved from

http://www.goleansixsigma.com/dmaic-five-basic-phases-of-lean-six-sigma/
 54

Guangdong Bureau of Statistics of China. (2013). The new registration transportation company

data. Retrieved from http://www.gdstats.gov.cn/tjgb/default.htm

He, Y., Tang, X., & Chang, W. (2010). Technical decomposition approach of critical to quality

characteristics for product design for Six Sigma. Quality & Reliability Engineering

International, 26(4), 325-339.

Harry, M. (1998). Six Sigma: A breakthrough strategy for profitability. Quality Progress, 31(2),

60-64.

Hahn, H., & Hill, W. (1999). The impact of six sigma improvement: A glimpse into the future of

statistics. The American Statistician, 53(2), 208-215.

Montgomery, D. C. (2009). Introduction to statistical quality control. Hoboken, NJ: John Wiley

& Sons.

Moore, J., & Weatherford, L. (2001). Decision modeling with Microsoft Excel. Upper Saddle

River, NJ: Prentice Hall.

National Bureau of Statistics of China. (2013a). Chinese GDP historical data. Retrieved from

http://www.stats.gov.cn/was5/web/search?channelid=288041&andsen=GDP

National Bureau of Statistics of China. (2013b). Chinese GDP historical data. Retrieved from

http://data.stats.gov.cn/search/keywordlist2;jsessionid=7AC6D61E4F733BDEEAC898F

E1B06E958?keyword=GDP

National Development and Reform Commission. (2012). Diesel prices historical data. Retrieved

from http://www.ndrc.gov.cn/hgjj/t20090114_256217.htm

Pande, P., Roland, N., & Cavanagh, R. (2000). The Six Sigma way: How GE, Motorola, and

other top companies are honing their performance. Hoboken, NJ: McGraw-Hill.
 55

Reytblat, B. (2011). New to Six Sigma. Retrieved from

http://www.isixsigma.com/new-to-six-sigma/

Ruan, Y. (2012). Chinese real estate investment increased significantly. Retrieved from

http://www.chinanews.com/estate/2011/08-09/3245733.shtml

Sahoo, P., Dash, R., & Nataraj, G. (2010). Infrastructure development and economic growth in

China, IDE Discussion Paper, 261, 68-70.

Snee, R. D. (1999). Why should statisticians pay attention to Six Sigma? Quality Progress,

32(9), 100.

Wei, X. (2012). Drivers purchase “Monthly Pass” for overloaded trucks. Retrieved from

http://news.qq.com/a/20120917/000430.htm
 56

Appendix A

ABC Transportation Company 2012 Shipping Cost Summary

Extra Rental Cost, Overload Fine, Extra Fuel Cost and Late Delivery Penalty in 2012

Month Extra Rental Overload Fine Extra Fuel Cost Late Penalty

RMB USD RMB USD RMB USD RMB USD

Jan-12 99,658 16,070 70,616 11,380 57,100 9,200 51,446 8,290

Feb-12 103,688 16,720 68,672 11,070 56,854 9,170 51,464 8,300

Mar-12 99,903 16,110 70,040 11,290 56,824 9,160 52,869 8,520

Apr-12 99,400 16,030 71,045 11,460 62,480 10,080 51,100 8,240

May-12 103,084 16,620 69,401 11,190 56,280 9,070 51,711 8,340

Jun-12 101,895 16,430 69,901 11,270 56,754 9,150 51,066 8,230

Jul-12 102,004 16,450 68,807 11,090 57,049 9,200 51,756 8,340

Aug-12 102,995 16,610 68,861 11,100 57,720 9,300 51,330 8,270

Sep-12 101,913 16,430 70,183 11,310 58,243 9,390 52,085 8,400

Oct-12 100,656 16,230 70,868 11,430 58,509 9,430 51,870 8,360

Nov-12 99,896 16,110 71,137 11,470 57,951 9,340 52,131 8,400

Dec-12 102,404 16,510 70,564 11,380 56,778 9,150 52,155 8,410

Total 1,217,496 196,370 840,095 135,490 692,542 111,700 620,983 100,150