An Architecture for Planning in Labour-Intensive Environments through Agent Modelling and Evolutionary Programming

In industry, typical scheduling systems rely on production rates which are, in general, highly static. This is the case for capital intensive enterprises but not the case in labour intensive situations. On the other hand, while most, if not all, researchers will agree that using Genetic and Evolutionary Programming (GEP) to achieve optimal solutions in Job-Shop Scheduling Problems (JSSPs) in not at all a new approach, most of the proposals to address the problem rely on purely non-symbolic systems. The introduction of Extended Logic Programming (ELP) provides a modeling basis for agents that emulate the workforce inside a company. Combining the expressiveness of ELP with GEP, labour intensive enterprises may now fully model their scheduling procedures. ELP is at the core of a new approach to Knowledge Representation (KR) and reasoning, favouring the application of a declarative language to KR in Artificial Intelligence (AI). It is considered an extension to the language of definite logic programs by classical negation, disjunction, meta-reasoning and reasoning in open domains, which extends the representation power of the language to handle incomplete information. This data is paramount for an effective search of the solution space in JSSP environments, being ignored in most of the non-symbolic systems presently available.