![]() | Only 14 pages are availabe for public view |
Abstract This thesis treated the assembly line balancing problem with genetic algorithms. The proposed model is capable for balancing single or mixed assembly line problems under the deterministic or stochastic task time. And a computerized program (ESAL) has been developed and coded with Visual Basic (VB 6). The assembly line balancing problem has proved out to have the characteristic of NP-hard, so the genetic algorithms (Gas) with the simple procedure that have the abilities of population based searching and parallel calculation are appropriate for solving the problem. To tackle a multi-objective problem, a weighted sum of objectives is used as a fitness function calculation to select the best solutions. The genetic algorithms used a feasible solution to the problem for building initial population chromosomes. Selection of the next generation is a combination between elitist models and tournament selection procedure with a tournament size of (2) as default. The crossover producer like partial mapping crossover are used with a crossover rate 0.95 as default, and random mutation with mutation rate equal to 0.04 as default was used, after the crossover and mutation procedure done the repair methods are created to make the produced off springs feasible solutions, all defaults values can be changed according to decision maker. The procedure of genetic algorithms is repeated until the termination condition was reached, which based on number of generations set to (500) generations as default value. In order to test the performance of the developed model, the suggested system and algorithms are applied to a wide set of different benchmark problems. Results showed the superiority of the suggested model and algorithms in terms of the quality of solution (the best known solutions are broken) and objective space exploration (a wider efficient set of solutions). The model solves all problems with known optimal number of stations and succeeds in some cases to reduce the cycle time and smoothness index. Finally, the developed system has been applied in an Egyptian Company for produces electrical transformers, the study shows that batch production is the common production method for this company, the mixed balancing can not applied unless a sequencing study done for the company. The results show how to increase the weighted efficiency for the batch production from 62.8% to 79.9% and to reduce weighted cycle time. In addition the weighted smoothness index will be reduced with about 54.8%. Reduced the number of workers, reduced the number of stations, and the probability of non stopping of the line will be attained. |