combined algorithms of optimal resource allocation结合最优资源分配的算法.pdfVIP

Applied Mathematics, 2012, 3, 78-85 /10.4236/am.2012.31013 Published Online January 2012 (http://www.SciRP.org/journal/am) Combined Algorithms of Optimal Resource Allocation Valery I. Struchenkov Moscow State Technical University of Radio Engineering, Electronics and Automation, Moscow, Russia Email: str1942@mail.ru Received October 17, 2011; revised November 19, 2011; accepted November 28, 2011 ABSTRACT Under study is the problem of optimum allocation of a resource. The following is proposed: the algorithm of dynamic programming in which on each step we only use the set of Pareto-optimal points, from which unpromising points are in addition excluded. For this purpose, initial approximations and bilateral prognostic evaluations of optimum are used. These evaluations are obtained by the method of branch and bound. A new algorithm “descent-ascent” is proposed to find upper and lower limits of the optimum. It repeatedly allows to increase the efficiency of the algorithm in the com- parison with the well known methods. The results of calculations are included. Keywords: Dynamic Programming; The Pareto Set; Branch and Bound Method; The Curse of Dimensionality; Algorithm “Descent-Ascent” 1. Introduction That is why the search of more effective algorithms es- pecially for the big dimension problem which built in The problem of optimum distribution of limited resource multiply repeated cycle of calculation is actual. This is R between n consumers was solved by R. Bellm