A coevolutionary genetic algorithm approach for multiperiod manufacturing planning problems with financial investment

We consider the multiperiod manufacturing planning problems with financial investment; that is, we consider two single-period plannings concurrently; One is the manufacturing planning to determine the number of kanbans and base stocks at each stage of the manufacturing system, and the other is the financial investment planning to determine portfolio under the investable surplus excluding a production fund from cash on hand. The manufacturing activities and the financial investment yield profits which are incorporated in cash on hand in the next period. Generally, it is hard to optimize these two plannings concurrently in finite time because of millions of combinations of feasible solutions. On the other hand, optimizing the manufacturing planning and financial investment planning independently leads to inaccurate solutions because of the trade-off between these two plannings. In this paper, we propose a new approach to obtain an approximate optimal combination with satisfactory accuracy using a coevolutionary genetic algorithm.