AI

Using Actors to Implement Sequential Simulations

Abstract

This thesis investigates using an approach based on the Actors paradigm for implementing a discrete event simulation system and comparing the results with more traditional approaches. The goal of this work is to determine if using Actors for sequential programming is viable. If Actors are viable for this type of programming, then it follows that they would be usable for general programming. One potential advantage of using Actors instead of traditional paradigms for general programming would be the elimination of a distinction between designing for a sequential environment and a concurrent/distributed one. Using Actors for general programming may also allow for a single implementation that can be deployed on both single core and multiple core systems. Most of the existing discussions about the Actors model focus on its strengths in distributed environments and its ability to scale with the amount of available computing resources. The chosen system for implementation is intentionally sequential to allow for examination of the behaviour of existing Actors implementations where managing concurrency complexity is not the primary task. Multiple implementations of the simulation system were built using different languages (C++, Erlang, and Java) and different paradigms, including traditional ones and Actors. These different implementations were compared quantitatively, based on their execution time, memory usage, and code complexity. The analysis of these comparisons indicates that for certain existing development environments, Erlang/OTP, following the Actors paradigm, produces a comparable or better implementation than traditional paradigms. Further research is suggested to solidify the validity of the results presented in this research and to extend their applicability.