ALGORITHM-744 - A STOCHASTIC ALGORITHM FOR GLOBAL OPTIMIZATION WITH CONSTRAINTS

A stochastic algorithm is presented for finding the global optimum of a function of n variables subject to general constraints. The algorith m is intended for moderate values of n, but it can accommodate objecti ve and constraint functions that are discontinuous and can take advant age of parallel processors. The performance of this algorithm is compa red to that of the Nelder-Mead Simplex algorithm and a Simulated Annea ling algorithm on a variety of nonlinear functions. In addition, one-, two-, four-, and eight-processor versions of the algorithm are compar ed using 64 of the nonlinear problems with constraints collected by Ho ck and Schittkowski [1981]. In general, the algorithm is more robust t han the Simplex algorithm, but computationally more expensive. The alg orithm appears to be as robust as the Simulated Annealing algorithm, b ut computationally cheaper. Issues discussed include algorithm speed a nd robustness, applicability to both computer and mathematical models, and parallel efficiency.