Resource allocation based on multi-objective programming in economy grid: design and performance evaluation

Abstract

This paper presents an economic-based resource allocation algorithm - mix optimization algorithm to meet user requirements in proportional resource sharing grid environment. The mix optimization algorithm is targeted to minimize two objectives, i.e., competing users 'processing cost and competing users' executing time with budget and deadline constraints, using multi-objective programming method. After establishing evaluating function of the optimal problem, a method similar to alpha-method is introduced to determine weight coefficients which define the relative importance of different optimization objective. The experimental results show that mix optimization algorithm leads to more efficient resource allocation than algorithms with single optimal goal (time optimization and cost optimization)