It is difficult to find optimal scheduling solutions for abstract scheduling problems with mass parallel tasks on multiprocessors because they are NP-complete. In this paper, a solution searching strategy called solution characteristic extraction is proposed as a multi-objective optimizer for solving flexible job shop problems (FJSP). These problems are concerned with finishing assigned jobs with minimal critical machine workload, total workload, and completion times. A suitable job assignment must consider processor performance, job complexity, and job suitability for each individual processor simultaneously. To test the efficiency and robustness of the proposed method, the experiments will contain two groups of benchmarks; with, and without release time constraints. Each benchmark includes numbers of heterogeneous processors and different jobs for execution. The results indicate the proposed method can find more potential solutions, and outperform related methods.
Digital Object Identifier (DOI)
Hsieh, Sheng-Ta; Yen, Shi-Jim; Lin, Chun-Ling; Chiu, Shih-Yuan; and Su, Tsan-Cheng
"Solving The Flexible Job Shop Problem using Multi-Objective Optimizer with Solution Characteristic Extraction,"
Applied Mathematics & Information Sciences: Vol. 10
, Article 23.
Available at: https://dc.naturalspublishing.com/amis/vol10/iss5/23