Benchmarking Communication in Actor- and Agent-Based Languages

This paper presents some results of communication benchmarks used to compare the performance of one agent-oriented and two actor-oriented programming languages. The experiments include an existing benchmark for traditional programming languages as well as two new variants of that benchmark. We selected Erlang and Scala to represent actor languages, and Jason to represent agent languages. We discuss here a summary of the result for those three experimental scenarios for each of the three languages and the respective result analysis in regards to time, memory, and core usage.

Authors

• Rafael C. Cardoso University of Aberdeen
• Jomi F. Hübner Federal University of Santa Catarina
• Rafael H. Bordini

Keywords

• communication benchmark
• multi-agent programming
• actor-based programming