Context Sensitive Dynamic Slicing of Multi-Agent System

Dynamic slicing is a method for finding and fixing bugs and testing various software systems based on a particular execution. In case of multi agent systems, we use agents as basic run-time entity to execute the functionality. The slicing criterion is used to identify the statements in a code that modify the values of variables at a specific location in a program, either directly or indirectly. Static slicing, on the other hand, considers all possible program executions instead of a specific execution. In this research work, we designed a method for dynamic slicing Multi-Agent Systems. The first step in our method is to generate a Multi-Agent System software execution trace, which records the order of events and interactions among agents during one execution. We use this trace to create a dependency graph that shows the connections between the statements in the program. Using this dependency graph, we apply a context-sensitive approach that takes into account the context in which statements are executed to produce more precise slices. Our aim is to provide complete debugging and testing methods for Multi-Agent Systems that allow developers to easily locate and isolate program parts that are related to a certain problem or behaviour.