改进缩张算法求解较复杂非线性规划问题

APPLYING IMPROVED C-E ALGORITHM TO SOLVE CONSTRAINED NLP

将缩张算法加以改进 ,较大程度地减少目标函数的访问次数 ,既可运用到规模和难度较大的约束非线性规划问题 ,也可运用到整数及混合规划问题。通过对 2个实例的数值分析 ,结果表明 :改进的缩张算法可实现实际问题的全局最优解 ,比其他优化算法在实现目标函数全局最优解上更加简便有效。

The Contraction expansion algorithm was improved by greatly reducing visiting objective function, so it could solve the large scale, integer and combined constrained nonlinear programming. The global optimal results were reached for two numerical examples. It clearly showed that improved C E algorithm was much better than other algorithms.