Parallel Robot Scheduling with Genetic Algorithms

摘要

The genetic algorithms (GA) have the great advantage and success in the solution of NP problems. There are various important applications on this way. In this study, the job with n-number of precedence constraints is assigned minimizing total earliness and tardiness and maximum flow time on m-number of parallel machine. Genetic algorithms and simulated annealing methods were used to find the solutions, which minimizes the total earliness and tardiness costs. In GA, the solution alternatives, which were obtained by using genetic operators, were investigated to understand that if they are feasible or not and the feasible ones according to precedence constraints were considered. The way, trying to make infeasible solutions feasible, was not selected. Likewise, obtained infeasible solutions were not evaluated. Again any study about making these infeasible solutions feasible was not done. According to the results obtained by using GA and SA methods, it was evidently observed that GA algorithm is more successful. Especially for larger problem sizes, it is seen that GA gives results better than SA.