| Summary: | 碩士 === 立德管理學院 === 應用資訊研究所 === 94 === This thesis proposed a Tabu-Genetic Algorithm (Tabu-GA), which incorporates a traditional Genetic Algorithm (GA) and Tabu Search (TS), to solve the global optimization problems. The principle of the proposed algorithm is that we introduce the Tabu Search to perform a fine tune in local search around neighbors on each individual within the population when an evolutionary generation of the GA is completed. Thus, the aim of the work is to find an acceptable optimal solution with less CPU time, and to enhance the poor efficiency in the local search for the traditional GA.
This study used global optimization functions with high dimensions for the efficiency evaluation of the proposed algorithm. The numerical results using the present algorithm were compared with those results obtained using the traditional GA. Furthermore, the proposed algorithm is applied to the production scheduling problem and compared the algorithmic capability with other methods. Based on the solutions solved the aforementioned benchmarking tests and scheduling problem, the Tabu-Genetic Algorithm shown that it is a useful method for solving the global optimization problem with high capability and efficiency.
|
|---|
