SA法在求解非线性约束优化问题中的应用

Application of SA algorithm to solving problem of non-linear constraint optimization

传统数学规划方法如梯度法等在解决非线性规划问题时,往往会由于问题本身的多峰性而落入局部最优解中,得不到全局最优解,这使得传统方法在解决非线性规划问题中受到很大的限制.80年代初,S.Kirkpatrick提出了模拟退火算法(Simutaneous Annealing),该方法在解决复杂的组合优化问题中可以得出很好的结果.它是一种仿金属退火物理过程的随机算法,在理想状态下可得出全局最优解,并能以一定的概率跳出局部最优解所在的区域.本文我们将探讨SA法在求解非线性约束优化问题中的应用.

Traditional mathematics programming methods met a olt of limitations in solving non-linear programming question because that question itself has multiple peak values. These methods such as grads algorithm always can figure out local optimization solution but not global optimization. In the beginning of 80's, S. Kirkpatrick presented Simultaneous Annealing algorithm. SA is a random algorithm of simulating metal annealing physical process and can get global optimization solution umder the ideal status, it can also jump out local optimization solution area in certain probability, so it is always used to solve complex compounding and optimization problem. In this paper, we will study SA algorithm in order to solve problem of non-linear constraint optimization.