A Three-Layer Approach to Testing of Multi-agent Systems

摘要

In this chapter, a complex approach to testing of multi-agent systems is presented. Multi-agent testing should bedivided into three layers. On the first layer, the individual agents should be tested. A stochastic approach to unit testingand the use of "stages" can be adopted. A stage is a perceptible segment of a virtual world where an agent is placed thatis projected for agent unit testing. On the second layer, deadlocks and similar flaws of agent interaction should befought. Our method of deadlock detection is based on monitoring recurring agent interactions. On the third layer, thebehavior of the entire system is evaluated. Bottlenecks and "hot spots" of the system could cause serious performanceproblems. Stability of multi-agent systems in case of mass collapse of many agents should be tested using stress tests.