Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 6

Chapter 5

RESULTS AND DISCUSSIONS

5.1 Introduction

The determination of suitable cutting parameters is one of the most


important elements of any metal part process planning. This work presents
a method for utilizing the Genetic Algorithm (GA) to optimize cutting
parameters in order to achieve the lowest unit manufacturing cost.

Parallel computing has made genetic algorithms even more appealing


by overcoming one of the main disadvantages: speed. This enables more
iterations and total runs, increasing the possibility of determining better
solutions and the certainty that the solutions discovered are optimal.

In this chapter, we will focus on running the genetic algorithm through


a series of iterations in order to find the best solutions for the research
problem. The algorithm will be implemented using programmed genetic
algorithm and the program is available in the MATLAB toolbox.
Subsequently, a new model will be utilized to validate the effectiveness of
the program in finding the optimal solution.

5.2 Results of genetic algorithm

We will use the same data in both comparisons in terms of number of


generations, population size, and elitism for (fitness function: @ mc-min;
bounds: lower [10 0.05 0.13] and upper [300 2.5 6.1]). Table 5.1 and Table
5.3 show the best result among all the 100 iterations and give the best
combination of the speed, feed, and depth of cut for the minimum
production cost.

Table 5.1 Output result of minimum production cost using my own


programming
Iteration Speed Feed Depth of cut Unit Cost
(no.) (m/min) (mm/rev) (mm) (UC)
1 172.08 0.3478 1.0003 0.87008
2 171.18 0.34856 1 0.86934
3 170.88 0.34794 1.0001 0.86992
4 173.56 0.3432 1.0004 0.87447
5 170.64 0.34483 1.0008 0.87291
6 173.95 0.34722 1.0001 0.87084
7 171.49 0.34795 1 0.8699
8 170.92 0.34701 1 0.87077
9 175.14 0.34782 1.0003 0.87061
10 172.79 0.34595 1.0001 0.87183

Table 5.2 indicates that the optimum production cost is 0.86934$ when
all of the values with optimized speed, feed, and depth of cut are used. We
see that the difference between the lowest and highest results obtained from
all 10 iterations was 0.005. This demonstrates the consistency of the
program's performance under the same operational settings.

Table 5.2 Optimized result of production cost


Speed Feed Depth of cut Unit Cost
(m/min) (mm/rev) (mm) (UC)
171.18 0.34856 1 0.86934

Table 5.3 Output result of minimum production cost using the MATLAB
toolbox
Iteration Speed Feed Depth of cut Unit Cost
(no.) (m/min) (mm/rev) (mm) (UC)
1 241.7 0.33345 1.0001 1.0497
2 201.62 0.34682 1.001 0.90275
3 65.587 0.34854 1 1.4079
4 378.38 0.1 1 1.9283
5 135.64 0.34834 1 0.91287
6 338.92 0.12789 1.0001 1.639
7 188.39 0.34831 1.0001 0.87988
8 167.25 0.34863 1 0.86973
9 201.92 0.34862 1 0.90227
10 147.63 0.34825 1.0001 0.88828

Table 5.4 indicates that the optimum production cost is 0.86973$ when
all of the values with optimized feed, speed, and depth of cut are used. We
see that the difference between the lowest and highest results obtained from
all 10 iterations was 1.04.

Table 5.4 Optimized result of production cost


Speed Feed Depth of cut Unit Cost
(m/min) (mm/rev) (mm) (UC)
167.25 0.34863 1 0.86973

This demonstrates that the program developed according to the method


used in the previous chapter's solution, which was run through a series of
iterations, produced results that were more converging to those obtained
from the MATLAB toolbox, as shown in Figure 5.1.
2

using my own programming


using the MATLAB toolbox

1.8

1.6
Cost

1.4

1.2

0.8
1 2 3 4 5 6 7 8 9 10
Iteration

Figure 5.1 Output result of minimum production cost

5.3 Model validation and evaluation of the effect of changing


generation size on cost value

We will use a new model to evaluate the correct performance of the


program and determine if the values between the program and the
MATLAB toolbox are comparable or close to each other, as happened
previously in the first model. The results of testing both models using a
number of iterations to find the lowest cost value from each generation are
shown in Table 5.5. The generations utilized to determine the size of the
effect of generation size on cost value are 50, 100, and 500.

The data for the model that will be utilized to validate the modeling
were obtained from (Cus and Bali 2003).
25.133 −3 0.7 0.55 1.22
f ( v , f , d )=0.0128+ +2.706 ×10 ×V × f × d
fV

Subject to the constraints: 70 ≤ v ≤100

0.1 ≤ f ≤1
0.1 ≤ d ≤ 5
1.18 1.26
1.38 f d ≤230
0.24 0.11
0.000626 f d V ≤5

Table 3 shows the optimum cutting


conditions obtained from PSO for
the pro
fi
le given in Fig.
3
and
shows that in the initial stages of the
iterations, PSO produces minimum
fi
tness and at the end of the
iterations, it gives better results by
using C++ language the optimal
solution can be obtaine
Table 5.5 Output result for minimum production cost for different
generations
by Genetic Algorithm (GA) by My Programing
toolbox

Depth of cut

Depth of cut
Generation

Unit Cost

Unit Cost
(mm/rev)

(mm/rev)
(m/min)

(m/min)
Speed

Speed
(mm)

(mm)
Feed

Feed
Models

50 211.6 0.3482 1 0.9268 171.05 0.348 1.0001 0.8696


Model
100 177.9 0.3486 1 0.8709 172.36 0.348 1 0.8695
(1)
500 174.2 0.3486 1.0001 0.8697 171.2 0.348 1 0.8693
50 99.37 1 0.1 0.2698 99.89 0.996 0.1088 0.2698
Model 100 100 1 0.1000 0.2682 99.99 0.999 0.1008 0.2683
(2)
500 99.99 1 0.1000 0.2682 99.99 0.999 0.1001 0.2682

You might also like