Cooperative Coevolutionary Optimization Method for Product Family Design

Under the mass customization paradigm, an increasing number of companies design a family of product variants simultaneously, instead of designing one product at a time. Product family design is a way to share resources through commonality among the products so as to shortening product lead-times while introduce adequate product variety in the competitive market in a cost-effective way. The key to using this approach successfully is to achieve the right trade-offs between commonality and performance product family. A product family design model is proposed as an optimization problem with two objectives. One represents the average performance of the product variants and the other is the measure of commonality of the product family in the presence of multiple levels. Optimal decisions include which design variables should be common among which product variants, and the values of each variable of each product variant in the family. A multiobjective cooperative coevolutionary algorithm is developed for simultaneous design of a family of product variants. Computational experiments are conducted using the design of a family of welded beams to demonstrate the effectiveness of the product family design method proposed in this paper.