المستخلص: |
This article presents general formulation of flexible job scheduling problem (FJSP) in industrial production. It describes genetic algorithm AGHAR that involves specific of technical-organizational questions in investigated production problem. It precisely represents chromosome structure and procedure of machines reservation. It also describes mutation and crossing operations and shows examples of its realizations. Genetic algorithm was worked out for serial and serial-parallel cases of parts flow. Its efficiency was compared to Szez heuristic algorithm. Our computational tests showed that this algorithm can improve solution of the FJSP. Evolutionary algorithm as a simulation of genetic process is not a random searching for solution of a problem. Although genetic algorithms use stochastic processes, their applications give generally better results than random. Obtained results of experimental analysis show that met heuristic methods, like genetic algorithms, are effective tool for solving problems very difficult from algorithmic point of view. It confirms that practical use of these methods is now one of the most effective ways of control in a real complex industrial processes. Popular methods for solving such type of problems presently are neural networks, tabu search algorithm, simulated annealing and genetic algorithms considered here. Efficiency of genetic algorithms is generally estimated by many criterions, like time of result return or quality of result.
|