基于离散粒子群算法的矩形件优化排样

Optimum packing of rectangles based on discrete particle swarm optimization

目前,粒子群算法在连续问题优化上的应用已经很广泛,然而在离散问题优化方面仍处在尝试阶段。提出了一种改进粒子群算法来解决矩形件排样优化问题(离散优化问题)。该算法融合了遗传算法中的交叉和变异思想,采用了信息交流策略,使其达到快速优化目的。算法也对"最低水平线法"解码方式进行了改进。实验结果表明,该算法具有快速、高效特点,与现有同类算法比较,在解决矩形件排样问题方面的优势明显。

At present, particle swarm optimization algorithm （PSO） has been applied to continued problems for several years, but it is still an attempt to use the algorithm in discrete problems. An improved particle swarm optimization algorithm is proposed to solve the packing problem of rectangles, which belongs to discrete problems. This algorithm combines the crossover and mutation in genetic algorithm and it is based on the information communication strategy in order to speed the optimization process. The mode of decoding based on ＂the lowest horizontal line＂ is also modified. The experimental results show that this algorithm is promising and more efficient than other algorithms in solving the packing problem of rectangles.