Guided Ejection Search for the Job Shop Scheduling Problem

摘要

We present a local search framework we term guided ejection search (GES) for solving the job shop scheduling problem (JSP). The main principle of GES is to always search for an incomplete solution from which some components are removed, subject to the constraint that a quality of the incomplete solution is better than that of the best (complete) solution found during the search. Moreover, the search is enhanced by a concept reminiscent of guided local search and problem-dependent local searches. The experimental results for the standard benchmarks for the JSP demonstrate that the suggested GES is robust and highly competitive with the state-of-the-art metaheuristics for the JSP.