Helping the Performance Evaluation of an Agent Run-time Framework: the SARL Experience Index

The performance evaluation of an agent platform is central in the agent-based modeling field. The agent platform and the hardware as well as the operating system modules, including any virtual machines, influence this performance. The impact of these hardware and operating system modules should be understood and evaluated due to their high impact on the global performances. In order to evaluate these two components, this paper proposes the scoring approach, named SARL Experience Index. The score is based on CPU, memory and disk sub-scores. We advocate this score may be helpful for the agent platform users for determining quickly the positive or negative impacts of a new deployment platform for an agent framework.