The capability reasoner is a tool we developed in order to enable multi-agent systems to optimize their performance with respect to how the agents are able to use one another's capabilities, and finally to enable these systems to solve a complex task by breaking it in small tasks which can be handled by individual agents

For this purpose, a capability description language was devised; it is a formal language and has been developed due to the necessity of having away to capture the features of real world capabilities

The elements of the language are:

  • Concepts - which map agent capabilities
  • Attributes - which describe the intrinsic characteristics of the capabilities
  • Relations - that link concepts together and to agents

 The functionality of the reasoner was built around the capability description language. This means that the reasoner supports all the features of this language, and is able to perform inferences using the mentioned language constructs as inputs

The inference part of the reasoning process is given by the fact that knowledge in the system is represented by the means of ontologies, which are able to capture semantics. The semantics add an extra dimension to the functionality of the system, since the barrier between syntax and semantics is broken. This provides the ability to take into consideration the meaning the database contents and to solve real problems using this data

Platform details

  • Environment: Microsoft Windows Vista/XP
  • Development tools: Sicstus Prolog
  • Requirements: Sicstus Prolog