Reuse by Inheritance in Agent Programming Languages

摘要

The need to modularise and thereby reuse complex agent programs has long been recognised in agent programming language research. Current approaches to agent modularity fall into two main categories: compositional; and environment-based. Motivated by a problem which requires two variants of the same agent to be built, this paper proposes a set of language extensions which add a complementary modularity style - reuse by inheritance - to agent programming languages. The extensions are designed so that they can be implemented in a preprocessor and added easily to an existing language, without affecting its type system, and with little or no change to its underlying interpreter.