API-Calculus

In computer science, Api-calculus was introduced in 2002 as an extension of pi-calculus to address some of the limitations of pi-calculus for modeling intelligent agents.^[1] More specifically, it addresses knowledge representation, organizational grouping and migration of agents among groups. Moreover, it has the potential for modeling the security aspects of agent-based systems.

Api-calculus introduces three new concepts over ordinary pi-calculus and its extensions, the higher order and polyadic pi-calculi. To represent knowledge inherent in an autonomous agent, the concept of a knowledge unit is introduced. A knowledge unit is an intelligence entity that can perform inference. Agents have the capability to add/drop facts (i.e. predicates or propositions) to/from a knowledge unit and also modify its structure by adding new rules or eliminating existing ones. Each mobile agent is capable of carrying one or more knowledge units and sending and receiving them to/from other agents. However, the concept of knowledge unit only provides an abstraction level with no resources for intelligence modeling. Moreover, api-calculus introduces milieu, a new level of abstraction that is in-between single mobile agents and the system as a whole. And lastly, Api-calculus introduces the notion of term. A term consists of a name, a rule/fact (used to create or modify knowledge units), or a function, where a name can be a channel or a variable.In the standard pi-calculus, names are the only terms.

References

1. ↑ (Rahimi 2002) Shahram Rahimi, Maria Cobb, Dia Ali, Fred Petry, “A Modeling Tool for Intelligent-Agent Based Systems: Api-Calculus,” Soft Computing Agents: A New Perspective for Dynamic Systems, the International Series "Frontiers in Artificial Intelligence and Application" by IOS Press, pp. 165-186, 2002.