A Performance Simulation Method of Concurrent System Based on Random Process Algebra

Abstract

The invention relates to a concurrent system performance simulation method based on random process algebra, which belongs to the technical field of computer applications. In the stochastic simulation technology of existing concurrent systems, it takes a long time to warm up the system from the initial state to the steady state, and this time-consuming will increase with the increase of the system scale. The method of the present invention utilizes the advantage that the Fluid approximation technology is fast and almost not affected by the system scale, quickly solves the approximate equilibrium state of the system, and allows the system to directly simulate from the vicinity of the equilibrium state to extract system performance parameters, eliminating the need for The time-consuming warm-up process can effectively improve the convergence speed of the algorithm and reduce the consumption of computing resources.

Description

technical field [0001] The invention belongs to the technical field of computer applications, and in particular relates to a concurrent system performance simulation method based on random process algebra. Background technique [0002] Concurrent systems in the fields of computer and communication often use random process algebra for protocol verification and performance evaluation [1,2]. Stochastic process algebra is a formal description language, which uses algebra as a tool to describe, analyze and reason about the behavior of concurrent systems (such as parallel computer systems, computer network systems, mobile communication systems, etc.). In this way, a system can be described as a group of processes that can perform actions, and these processes interact through the synchronization of actions. Each action is given a random variable describing the execution time of the action, usually these random variables obey the exponential distribution, so every algebraic model o...

AI Technical Summary

Problems solved by technology

[0022] In view of the stochastic simulation method used in the performance extraction technology of the existing concurrent system, which has the defects of slow convergence speed and long time consumption, the purpose of the present invention is to provide an improved stochastic simulation technology, which can greatly shorten the system running time from the initial state. Warm up to the equilibrium state to quickly obtain the performance parameters of the concurrent system running in the equilibrium state

Images