A Raindrop Algorithm for Searching The Global Optimal Solution in Non-linear Programming

摘要

In this paper, we apply the random walk model in designing a raindropalgorithm to find the global optimal solution of a non-linear programmingproblem. The raindrop algorithm does not require the information of the firstor second order derivatives of the object function. Hence it is a directmethod. We investigate the properties of raindrop algorithm. Besides, we applythe raindrop algorithm to solve a non-linear optimization problem, where theobject function is highly irregular (neither convex nor concave). And theglobal optimal solution can be found with small number of iterations.