Job shop rescheduling by using multi-objective genetic algorithm

Xinchang Hao,Lin Lin
DOI: https://doi.org/10.1109/ICCIE.2010.5668422
2010-01-01
Abstract:In current manufacturing systems, production is a very dynamic process with many unexpected events and continuously emerging new requirements. Researchers have developed a wide variety of procedures and heuristics for solving these scheduling problems, called rescheduling. Most proposed approaches are often derived by making simplifying assumptions. As a consequence, the approach is not in accordance with functioning of the real manufacturing system. Such approaches are frequently not suitable and flexible enough to respond efficiently to fast changes in the environment. In this paper, we focus on a practical solution of rescheduling by using mathematical modeling and interactive adaptive-weight evolutionary algorithm. We extend the rescheduling problem to a multi-objective optimization model. We formulate several objectives for corresponding requirement, such as due date, capability, transportation cost, set up cost and available resources etc. We can select the necessary one objective or some objectives for the manufacturing flexibility. However, for traditional approaches of multi-objective optimization problems, researchers focused on the solutions diversity. For the multi-objective rescheduling problem (moJSRS), we have to consider not only the solutions diversity, but also adapting the objectives alternative. We will propose an interactive adaptive-weight evolutionary algorithm with adapting the characteristics of a multi-objective job shop rescheduling problem. Some practical test instances will be demonstrated the effectiveness and efficiency of the proposed algorithm.
What problem does this paper attempt to address?