Smart Caching Algorithm for Software Agents Based on Re-execution Probability

  • Authors:

  • Jarogniew Rykowski

  • Affiliations:

  • The Poznan University of Economics, Department of Information Technology, 60-854 Poznan, Poland

  • Venue:
  • HoloMAS '07 Proceedings of the 3rd international conference on Industrial Applications of Holonic and Multi-Agent Systems: Holonic and Multi-Agent Systems for Manufacturing
  • Year:
  • 2007

Quantified Score

Hi-index 0.01

Visualization

Abstract

In this paper we propose a new, smart strategy for efficient caching of user-defined software agents. The strategy is based on setting up minimum and maximum caching time for individual agents. These timings are formed on the basis of a way of calling an agent (type of input/output gateway serving a call, previous calls to the same agent, etc.), on the one hand, and past/current/ predicted system load, including waiting calls to be served in the near future for the same and other agents, on the other hand. As proven by the tests in real applications, the strategy is efficient, especially in the case of a mass-scale system with thousands and even millions of agents owned (and possibly programmed) by different users.