Mathematics · Numerical Analysis
Greedy Algorithm with Relaxation in a Multiclass Queueing Environment
Wai Ki Ching, Associate Member
See on the mapAbstract
Abstract—In this letter, we consider the convergence of asyn-chronous greedy algorithm with relaxation for the Nash equi-librium in a noncooperative multiclass queuing environment. The process of asynchronous greedy algorithm is equivalent to the iteration of Jacobi method in solving a linear system. However, it has been proved that the algorithm converges only for some particular range of queuing parameters. Here we propose the asynchronous greedy algorithm with relaxation, which is in principle equivalent to solving a linear system by Jacobi method with relaxation. We propose also some relaxation parameters such that our algorithm converges very fast. Index Terms—Asynchronous greedy algorithm, Jacobi method, Nash equilibrium. I.