TABU SEARCH AND GENETIC ALGORITHM FOR PRODUCTION PROCESS SCHEDULING PROBLEM

Background: The paper deals with production process scheduling problem. In large companies, the decision-making process about operators’ work, machines availability and production flow is a very difficult task, which is often being done by employees. Thus, not always the decision made is optimal in terms of cost, production time, etc.

Methods: As a solution, two intelligent methods: Tabu Search and the genetic algorithm have been analyzed in field of production scheduling. The aim of this work was to examine the possibility of improving presented decision-making process that is being performed when scheduling, using Tabu Search and genetic algorithms. As a result of experimental re-search, it has been confirmed that the use of appropriately selected and parameterized intelligent methods allows for the optimization of the analyzed production process due to its du-ration. The research was case of study performed in cooperation with company that produces components for automotive industry.

Results: Basing on collected and analyzed data, considered methods can be more or less successfully used in production process scheduling. Comparing both used algorithms, Tabu Search twice proposed worse solutions, the average operational time was 1.63% shorter than the actual one. In this case, better results were reached by using genetic algorithm – potential operational time was always shorter than the actual one, and it was reduced by 6.3% in total on average.

Conclusion: Using algorithms allowed to achieve lower workload of employees and to reduce of operational time, which were the evaluation criteria in performed research. Managers of the analyzed company were pleased with the proposed solution and declared interest in developing these methods for future. This shows that intelligent methods can find, in relatively short time, the solution that is close to the optimal and acceptable from the problem point of view.