"Scalability Analysis of the Zereal Massively Multiplayer Game Simulator"

Authors: Amund Tveit

Abstract:
This paper presents a mixed-level factorial design for scalability analysis of the Zereal Massively Multiplayer Game Simulator. The selected factors are number of killers, monsters and CPUs in addition to the killer and monster vision radius. The analysis show that the factors explain approximately 85% of the model (measured in square of errors), which is surprisingly good since the simulation is highly stochastic due to the 1-step Markov based intelligence of monsters and killers.

Keywords: Factorial design and analysis, Massively Multiplayer Games, Simulation, Performance

Bibtex Citation:

@TechReport{2002:TR:Tveit,
  author = 	 {Amund Tveit},
  title = 	 {Scalability Analysis of the Zereal Massively Multiplayer Game Simulator},
  institution =  {Department of Computer and Information Science, NTNU},
  year = 	 {2002},
  month = 	 {November},
  ISSN =         {0802-6394},
  number = 	 {12/02},
  address = 	 {Sem Sælands vei 7-9, NO-7491 Trondheim, Norway}
}

1. P. Quax, P. Monsieurs, T. Jehaes, W. Lamotte and U. Campus. "Using autonomous avatars to simulate a large-scale multi-user networked virtual environment", Proceedings of the 2004 ACM SIGGRAPH International Conference, USA, 2004

2. Håvard Engum, Jørgen Vinne Iversen, and Øyvind Rein. "Zereal: A semi-realistic simulator of Massively Multiplayer Games" Technical Report, NTNU, Trondheim, Norway ,November, 2002