【摘要】 针对智能优化算法求解大规模矩形排样问题时，求解速度慢、稳定性差的弊端，提出了一种启发式算法，即带对偶项的降序有限首次适应算法。经实验验证，算法在测试集上的求解时长在100 ms内，求解出的板材利用率可达95%左右，相较于现有启发式算法有2%的提升。此外，针对个性化订单矩形件材料强互异性的情况，提出了一种考虑材料相似度的二阶段贪心算法。定量描述订单间材料的相似度，最大程度的将相似度高的订单组批进同一生产批次来提高板材利用率。实验表明，用该算法对订单进行组批优化后再进行排样求解出的板材利用率相较于未考虑材料相似度有12%的提升。更多还原

【Abstract】 Aiming at the disadvantages of slow solution speed and poor stability when intelligent optimization algorithm solves large-scale rectangular layout problem, a descending finite first fit heuristic algorithm with dual term is proposed.It has been verified by experiments that the solution time of the algorithm on the test set is within 100 milliseconds, and the utilization rate of the plate can reach about 95%,which is 2% higher than the existing heuristic algorithm.In addition, a two-stage greedy algorithm considering the similarity of materials is proposed for the case of strong heterogeneity of materials for personalized order rectangular parts.Quantitatively describe the similarity of materials between orders, and batch the orders with high similarity into the same production batch to the greatest extent to improve the utilization rate of plates.Experiments show that using this algorithm to group and optimize orders, and then carry out layout to solve the board utilization rate has a 12% increase.更多还原