A MODELING OF INTERACTIVE FACILITIES LAYOUT DESIGNER REASONING USING QUALITATIVE PATTERNS

A modelling of interactive facilities layout designer reasoning for the purpose of improving the quality of solution currently available in the layout literature. Some of the constraints which are laborious to explore interactively, are enumerated by automated reasoning. The facilities layout design task is conceived as a multi-agent collaborative reasoning problem. The described approach is fundamentally different from previous approaches in three ways: (i) The reasoning is focused on the automated identification and rectification attempts on a set of qualitative patterns (termed qualitative layout anomalies or QLAs). (ii) The problem is solved by context-based reasoning, with each context originating from a different starting solution. The contexts allow one to jump into interesting regions of the solution space. (iii) An automated interface of the reasoning process and linear optimization is developed to solve the layout problem. This makes it more convenient to enumerate the solution space more exhaustively than is typically possible by a comparable ad hoc human interactive layout reasoning coupled with linear optimization. The methodology is implemented in an object-oriented environment using Smalltalk-80. A few experimental results are given and these results are compared with previous results on similar cases.