Solving multi-objective semiconductor assembly and test manufacturing scheduling problem based on estimation of distribution algorithm

摘要

In order to solve the problem that, from the current practical perspective of semiconductor assembly and test manufacturing (ATM), the traditional multi-objective optimization algorithm is difficult to realize the practical decision of the enterprise, this paper proposes a new multi-objective estimation of distribution algorithm (EDA) to solve the ATM scheduling problem. According to the demand of the ATM enterprise, based on sub-module using two modeling ideas, objectives were divided into two categories: constrained objective and optimized objective, and the different objective had the different searching process. Finally, it used the new algorithm to solve the multi-objective ATM scheduling problem. The result shows that the novel algorithm has the good feasibility and it also has an obvious advantage, the better practicability and maneuverability, compared with the traditional multi-objective optimization methods.