Study on multi-objective search performance of production plan optimization in manufacturing field

Abstract

There is a great demand for multi-objective optimization of production plans at the high-mix low-volume production of the manufacturing industry, and the importance of multi-objective numerical optimization technology will increase more and more in the future. In this research, we have developed a control method that can realize a wide range of Pareto solution at high speed for two object functions; cost and productivity, which are important in manufacturing fields but have a trade-off relationship. The key of present technique is a technology that can obtain a wide range of Pareto solutions, which are difficult to solve by each technology alone, by appropriately controlling a technology that has high search performance but is single objective optimization and a technology that can perform multi-objective optimization but has low search performance. This paper presents an example of solving a standard multi-objective optimization problem and the results of applying it to setup optimization problems which are often found in actual manufacturing fields.