A fuzzy based algorithm to solve the machine-loading problems of a FMS and its neuro fuzzy Petri net model

This paper aims to develop a fuzzy-based solution approach to address a machine-loading problem of a flexible manufacturing system (FMS). The proposed solution methodology effectively deals with all the three main constituents of a machine loading problem, viz. job sequence determination, operation machine allocation and the reallocation of jobs. The main objectives of the FMS loading problem considered here are minimisation of system imbalance and maximisation of throughput; the constraints to be satisfied are the available machining time and tool slots. An analytical argument has been provided to support the membership function related to the operation machine allocation vector. Computational results revealed the superiority of the proposed algorithm over other heuristics when it is tested on a standard data set adopted from literature. A new class of petri net model called the "Extended neuro fuzzy petri net" is constructed to capture clearly the various details of the machine loading problem which can be further extended to learn from experience and perform inferences so that truly intelligent system characteristics can be realised.