A fuzzy linguistic approach for operator scheduling problem under uncertain operatin times

This study presents an application of parallel processor scheduling under uncertain operation times. In most of manufacturing processes, the actual processing times become known after the completion of the operations, and the problem of the decision maker is to create schedules which suit real life. Operation times change because of some environmental effects, mostly because of operators. Operator skills, batch sizes, learning effect are the parameters that have been used as linguistic variables and a linguistic reasoning approach has been proposed in the study. We are motivated from a real case scheduling problem that contains some uncommon welding operations to be processed by workers in an automotive subcontract company. Here each operator may weld each part but in different processing times. The problem modeled as non-identical parallel processor scheduling problem under uncertainty and a fuzzy linguistic approach is aimed for solving. The developed approach consists of a Mamdani inference system with a 75-"If-Then" rules that determine crisp operation times from fuzzy values and "longest processing time" (LPT) heuristic algorithm has been used. The scheduling objective is to balance the workload among all operators. To comparison of efficiency of the method, the schedules that use random operation times generated from the uniform distribution were prepared and a simulation experiment was been done for 1000 times. Results showed that the proposed approach has an important contribution to solve the non-identical parallel processor scheduling problem under uncertainty.