Modelling and Optimizing for Distributed Systems with limited Resource Migration Capacity

Abstract

Abstract This paper discusses the resource scheduling of some distributed systems, such as logistics system and the mobile system. By analysing the characteristic of the resource allocation, we found the resource or load of the distributed systems has limited migration capacity, but the load can be transmitted between neighbour nodes. By diverting resource or load between neighbour nodes, the system can reach a more balanced state. Then a model, shown as a constrained optimization problem, is presented to describe the load balancing problem of the distributed systems. The constrained terms are processed by the penalty function, and then an algorithm based on penalty function is presented to obtain an optimal solution of the constrained optimization problem. Finally, we design some experiments to verify the load model and the optimization algorithm proposed in this paper.