产品协同设计系统中交互式双边协商问题研究

为了解决产品协同设计系统中的冲突消解问题，提出一种基于多目标遗传算法的交互式双边协商算法。该算法通过目标加权法将向量优化问题转化为标量优化问题，其中不同的权重向量可以提供不同的Pareto解。在该算法实现过程中，采用实数编码法进行编码，每个染色体代表一组描述协商方案的属性集。遗传操作包括选择、交叉和变异，其中方案的选择操作采用排序法，避免了在群体规模固定的情况下产生过量的后代。提出了一种最优保存策略，让方案群中适应度最大的方案不参与交叉和变异，而直接去替换适应度最小的方案，提高了算法的求解速度。最后，通过一个成本控制问题进行了检验，证明该算法不仅能得到协商结果，而且可以得到更多的Pareto解，为协同设计系统提供更多的选择方案和选择空间。

Research on interactive bilateral negotiation algorithms in product collaborative design system

An interactive bilateral negotiation algorithm based on multi-object genetic algorithm was presented to resolve the conflicts in product collaborative design system. Object-weighted method was used to transform the vector optimization problems to scalar quantity optimization problems and among which different weight vector could offer different Pareto solutions. In the implementation process of algorithm, a real number coding method was adopted with each chromosome representing an attribute set describing the negotiation project. Genetic operations included selection, crossover and variation. Sorting method was employed in selection operation to avoid producing superfluous offspring in the condition of fixed scale of colony. An optimum saving strategy was presented, which enabled the project with the biggest fitness in the project set to replace the project with the smallest fitness without conducting the operation of crossover and variation therefore the problem-resolving speed was improved. Finally a cost-control problem was put forward to verify this algorithm, and the result indicated that not only negotiation results but also more Pareto solutions could be obtained by the algorithm. It has supplied more selection projects and selection space in collaborative design system.