Abstract: Aiming at the multi-components layout problem with equilibrium constraints,an interior point method with global optimization ability was proposed.The local optimal result was searched by using an interior point algorithm which combined line search and trust region steps.Further,a special cross strategy which was used to change the layout pattern was proposed to enable the global search ability.Three classical layout problems were used to demonstrate the optimizing ability of proposed method.By comparing with some main algorithms,further analysis was carried out.Numerical results showed that the proposed method had close result and efficiency on small-scale instance and had the best result on large-scale instance.

Key words: multi-components, layout optimization, packing problem, interior point method, self-cross strategy

摘要： 针对带性能约束的圆形多组件装填布局问题,提出一种具有全局寻优能力的内点法。该算法采用线性搜索和信赖域方法相结合的内点法进行局部寻优;为使算法具有全局寻优能力,提出了自交叉操作的布局变换策略,从而跳出局部最优布局模式,实现组件装填布局的优化求解。对3个经典算例进行了计算验证,并重点与目前的主流算法进行深入分析对比。算例结果表明,该算法在小规模布局问题上的计算效率和精度与其他算法相当,对于规模较大的布局问题则有更好的计算结果,刷新了目前的最优结果。

关键词: 多组件, 布局优化, 装填问题, 内点法, 自交叉