Logistics transportation problems with linear programming
6.3 Literature review
6.3.1 Linear programming LP was initially used, along with many other management science techniques, as a discipline in the 1940s during and shortly after World War II Its development quickly accelerated after World War II as many industries found the technique valuable and began adopting it as a standard tool to allocate resources in an optimal way. LP is categorized under tools and techniques applied in decision analysis in OR where the problem data are known deterministically. Deterministic and probabilistic decision models are part of decision-making problems.Optimizing system performance in TPs necessitates continual improvement to minimize cost and maximize performance. A preferred method is to use a model accurately representing the appropriate aspect of the operation process to describe potential solutions. A mathematical optimization model consists of an objective function and a set of constraints expressed in the form of a system of equations or inequalities. Today, LP is used to solve decision problems in many interesting and important applications that may contain thousands of variables. Some of applications for its use include: ● Product mix planning ● Distribution networks ● Truck routing ● Staff scheduling ● Financial portfolios ● Corporate restructuring 6.3.2 Transportation problems The TP (also known as the transportation theory) was formalized by the French mathematician Gaspard Monge in 1781. During World War II, major advances on the theory were made by the Soviet mathematician and economist Leonid Kantrovich. 6.3.2.1 Transportation problem defined The main purpose of the TP is minimize the cost of satisfying the needs of each demand center and operate within the capacity of the supply source. Supply sources may sometimes be referred to as “origin.” Each supply source has a predetermined supply capacity and each demand destination has a predetermined level of demand that has to be satisfied. The objective of a TP is to ensure that the correct number of products is shipped between each source and destination while minimizing shipping cost and meeting demand and supply constraints by not exceeding supply at each source but meet demand requirements of destinations. Fig. 6.1 is an Illustration of supply to demand shipment. The TP solution’s key objective is to reduce the cost of the transportation. There are many methods currently being employed to solve TPs including computer-aided methods. The majority of the presently used methods for solving TPs are trying to reach the optimal solution. 6.3.2.2 Importance of transportation problem and linear programming in business and industry Today, LP is used to solve decision problems in many interesting and important applications that may contain thousands of variables. Some of the applications for its use include: ● product mix planning ● distribution networks ● truck routing ● staff scheduling ● financial portfolios ● corporate restructuring In business and industry today, LP is so effective it is said to account for as much as 90% of all computing time for business management decisions. In a relatively short period of time, it has changed the way business managers make decisions, from that of guesswork and intuition, to the use of an algorithm based on available data that accurately produces optimal decisions. Transportation is vital to the sustainability of many businesses and hugely affects the economy of many countries. This is mainly because businesses basically thrive on trading and transportation is the conveyor of any form of trade (goods or services). The geographical spread of the human population and natural resources and the need for trading among different countries means transportation plays a critical role in many economies. 6.3.2.3 Types of transportation problems The TP can be balanced or unbalanced. A balanced TP and an unbalanced TP are discussed as follows; Balanced TP: When the total amount demanded at all destinations is equal to the total supply available at all sources. Unbalanced TP: When the total demanded at all destinations is not equal to the total supply available at all sources. Even if the TP is unbalanced, it is possible to balance the problem. To balance an unbalanced TP, you need to add a dummy line. A dummy line is any arbitrary destination or source which is introduced into the problem such that it does not change the problem. If your total supply is less than your total demand, you add a dummy source and let the supply equal the difference. The shipping cost in a dummy destination or shipping cost from a dummy source is always zero. The TP can be visualized as a set of nodes representing m sources and n destinations interconnected by direct “arcs,” which represent the routes from each source to a destination (Fig. 6.2). The total required number of direct arcs is equal to m X n 6.4 Solving transportation problems 6.4.1 LP Formulation Suppose a company has m warehouses, i (source or origin) and n retail outlets, and j (demand center). Products are to be shipped from the warehouses to the retail outlets. Each warehouse has a given level of supply, and each retailreatil outlet has a given level of demand. Also, we are given the transportation cost between every warehouse and the retail outlet, and these costs are assumed to be linear. More explicitly, the assumptions are: The total supply of the products from warehouse i = ai, where i = 1,2,3...m. The total demand of the products at the retail outlet j = bj, where j = 1,2,3,...n. The cost of sending one unit of the product from warehouse i to retail outlet j is equal to Cij, where i = 1,2,3...m and j = 1,2,3,....n.
6.4.1.1 Decision variables
The variables in the LP model of the TP will hold the values for the number of units shipped from one source to a destination. The decision variables are: Xij = the size of shipment from warehouse i to outlet j, where i = 1,2,3...m and j = 1,2,3,...n. The number of decision variables = m X n. 6.4.1.2 Objective function The objective function contains the costs associated with each of the variables. It is a minimization problem. Consider the shipment from warehouse i to outlet j. For any i and j, the transportation cost per unit Cij and the size of the shipment is Xij. Since we assume that the total cost function is linear, the total cost of this shipment is given by CijXij. Summing over all i and j now yields the overall transportation cost for all warehouseoutlet combinations. Our objective function is then 6.4.1.3 Constraints TPs pose conditions that have to be satisfied to be able to obtain a true result. These conditions are known as constraints. In a TP, every node (each source or destination point) has a constraint. Hence, the total number of constraints for a TP equals m+n. Let ai be a source capacity and bjdenote destination needs. 6.4.2 Steps in solving a transportation problems In order to solve a TP, you must first understand the necessary steps it takes to solve the problem. There are four common steps used to solve a TP: 1. Define the problem and prepare the transportation tableau. 2. Obtain an initial feasible solution. Logistics transportation problems with linear programming 3. Identify the optimal solution. 4. Understand special situations. 6.4.2.1 Define the problem and prepare the transportation tableau The transportation tableau clearly articulates the supply and demand constraints and the shipping cost between each demand and supply point. Table 6.1 is an example of a transportation tableau. 6.4.2.2 Obtain an initial feasible solution There are various initial basic feasible solutions used to solve a TP. Here are a few of commonly used methods. 1. North-west corner method 2. Least-cost method 3. Vogel approximation 6.4.2.2.1 North-West corner method The north-west corner method is a method for finding the initial basic feasible solution where the basic variables are selected from the top-left corner (north-west of the table). Even though it may be simple and easy to use it often produces an insufficient solution, which can be considered a disadvantage. 6.4.2.2.2 Least-cost method In this method, the basic variables are chosen according to the transportation cost. The solution is found by considering the cheapest cell, assigning most of the supply to that cell, crossing out the figures in the corresponding rows and columns and modifying the figures of the remaining cells accordingly. 6.4.2.2.3 Vogel’s approximation method VAM is an improved version of the least-cost method that generally, but not always, produces better starting solutions. VAM is based on the concept of minimizing opportunity (or penalty) costs. The opportunity cost for a given supply row or demand column is defined as the difference between the lowest cost and the next lowest-cost alternative. This method is preferred over the methods discussed above because it generally yields an optimum, or close to optimum, starting solution. Consequently, if we use the initial solution obtained by VAM and proceed to solve for the optimum solution, the amount of time required to arrive at the optimum solution is greatly reduced (Asase, 2013)
6.4.2.3 Identify the optimal solution
This step seeks to make improvements to the initial feasible until no further reduced cost of transportation is possible. To ensure this, every unoccupied cell in the transportation tableau is analyzed. The method widely used for this stage is the stepping stone method using modified distribution (MODI) 6.4.2.4 Demand not equal to supply (unbalanced problem) A situation occurring frequently in real-world problems is the case where total demand is not equal to total supply. These unbalanced problems can be handled easily by the solution procedures discussed above if we first introduce dummy sources or dummy destinations. In the event that total supply is greater than total demand, a dummy destination, with demand exactly equal to the surplus, is created. If total demand is greater than total supply, we introduce a dummy source (factory) with a supply equal to the excess of demand over supply. In either ease, cost coefficients of zero are assigned to each dummy location. 6.5 Graphical method The TP can be easily solved by the graphical method if the problem has only two decision variables. By applying the graphical method, we are able to easily see the feasible region and determine the optimal solution. The feasible region represents a set of solution space that is satisfied by all of the constraints. Any point within the feasible region can be a solution of that problem. The graphical method cannot be used when there are more than two decision variables. The following steps are involved: 1. Determine the objective of the problem (in the TP, we already know the objective is to minimize transportation cost). 2. Find and simplify the constraints. 3. Plot the graph and determine the feasible area. 4. List all the vertices of the shaded area. 5. Evaluate the objective function Z =ax +by for each of the listed vertices. 6. Determine the minimum value from the evaluation done in the previous step