全局优化的一种新方法

A New Method for Global Optimization

近年来 ,组合优化领域中出现的新型随机搜索方法———蚂蚁算法 ,在著名的货郎问题 (TSP)以及一系列离散优化问题中获得成功并表现出相当好的性能。在这种生物算法思想的基础上进一步加以引深和扩展 ,提出了一种求解一般无约束函数优化问题的新方法 ,并给出其全局渐近收敛性 ,从实验上求解了一系列典型的测试函数 ,收到了良好的效果。由于无约束函数优化是一般全局优化问题中的基本情形 ,因此 ,该方法亦可为有约束情形下的函数优化提供了新的求解手段。

Recently, a new stochastic searching algorithm——ant algorithm, has been raised in combinatorial optimization area. Successful applications in the famous travelling salesman problem and several discrete optimization problems have shown its good properties. Based on the main idea of this biological algorithm, this paper further extends its usage and presents a new method for solving general unconstrained function optimization problems. Global convergence is offered and series of typical test functions are solved experimentally which give promising results. Since the unconstrained function optimization is the basic situation of general global optimization problem, this method can also be served as a new tool for function optimizaiton in constrained situation.