基于零件相似性特征的大规模下料分组优化方法

针对诸多优化算法在处理大规模下料问题时易于陷入局部最优解和时间效率低下的问题,提出基于零件相似性特征的分组优化方法.该方法利用HCM算法进行零件聚类,产生若干相似组,然后通过分析计算机硬件水平和算法处理问题能力将相似组重组为若干分组,再对各分组依次优化.在相邻分组的优化中,采用补偿策略动态修正零件在分组间的分布,优化完毕后合并各组优化结果得到原问题的下料方案.实验结果表明,相对于不分组优化,该方法处理大规模下料问题的效率高,且可以维持和提高材料利用率.

A Grouping Optimization Method for Solving Large-Scale Cutting Stock Problem Based on the Similarity of Parts

A new grouping optimization method based on the similarity of parts is proposed,which is intended to solve the low-efficiency and local optimization problems in large-scale cutting stock problem(LCSP).By using HCM algorithm,all the parts are clustered at first.Then,the part-clusters are recombined into several part-groups by considering the computer hardware conditions and the optimization algorithm's characteristics.The LCSP is decomposed into several small-scale cutting stock problem(SCSP).After that,all the SCSPs are solved separately.For every pair of two adjacent SCSPs,a compensation strategy is adopted to adjust parts in different groups.Eventually,the result of the LCSP is obtained by combining all the results of the SCSPs.Compared with the general optimization algorithms,the proposed method is highly effective both in time-efficiency and utilization ratio of materials.