Debugging and testing of multi-agent systems using design artefacts

Agents are a promising technology for dealing with increasingly complex system development. An agent may have many ways of achieving a given task, and it selects the most appropriate way of dealing with a given task based on the context. Although this makes agents flexible and robust, it makes testing and debugging of agent systems challenging. This chapter presents two tools: one for generating test cases for unit testing agent systems, and one for debugging agent systems by monitoring a running system. Both tools are based on the thesis that design artefacts can be valuable resources in testing and debugging. An empirical evaluation that was performed with the debugging tool showed that the debugging tool was useful to developers, providing a significant improvement in the number of bugs that were fixed, and in the amount of time taken.