Improvement And Performance Analysis Of Stochastic Simulation Algorithm Of A Class Of PEPA Models

The process algebra is a modelling formalism.Based on algebra, it is a method which can be used to describe, analyse and deduce the synchronisation systems. In1990s, the process algebra developed and became the stochastic process algebra under the condition that every action was given a stochastic variable which was used to represent the time that the action was performed. The stochastic process algebra PEPA is a powerful modelling formalism, which has enjoyed successful applications in performance modelling of computer and communication systems. However, the state-space explosion problem will result from a large scale system due to its large numbers of components. Both quantitative and qualitative analysis of large model systems suffer from this problem, which has restricted wider application of the PEPA language. Thus, this paper introduce a more perfect numerical representation scheme and stochastic simulation algorithm which is based on numerical representation can simulate a larger scale system and give the performance measures which are needed. However, due to the more and more components, the response time (the time that the systems work from the initial state to equilibrium state) of the stochastic simulation process become longer and longer, thus the increasing resources will be consumed. This paper presents a simple PEPA model, the system which combines the fluid approximation approach is going to simulate directly from the nearby state of the equilibrium state instead of initial state which can dramatically reduce the response time from the initial state to equilibrium state of the simulation process.In Chapter1, the background, history about the related work and the origin of our question are first introduced and the major work of this paper is presented.Chapter2is devoted to introducing PEPA, explaining the state-space explosion problem and presenting the numerical representation scheme and the fluid approximation approach and stochastic simulation algorithm based on numeric representation.Chapter3will present a simple PEPA model and numeric solution which is resulted from a set of ordinal differential equations (ODEs) and seek out a state which approaches the solution from the state space.In Chapter4,the corresponding numerical simulations are given by Matlab to respectively illustrate the performance measures between the initial algorithm and improved algorithm and give the performance analysis.Finally the Chapter5will conclude the paper and propose future work.