Multi-Objective Flexible Job Shop Scheduling Optimization Models (I)

In manufacturing area, the most difficult problems related to the time-efficiency frequently occur if the jobs to be scheduled are heterogeneous, if there are alternative routes on the machines for the jobs and multiple objectives are imposed. This framework, named Multi-Objective Flexible Job Shop Scheduling Problems (MOFJSSP), has been formalized, modeled and studied in the last decades from many perspectives in order to capture all the influential factors and to optimally solve the multi-objective problem while satisfying the multiple specific constraints. The author focused the research on the various optimization models adequate for the (MOF)JSSP, both conventional and non-conventional: Petri nets, waiting systems, general decision models, logical formulations as STRIPS language, Markov processes, Monte Carlo simulation and procedural models (agent-based models, evolutionary algorithms, expert systems, fuzzy techniques, neural networks). The structure, classification and analysis of the optimization models is extensively new, as the MOFJSSP models were less reported in the literature than the JSSP models. All the models are analyzed in the research (which is divided into two parts), in general case and for a specific real JSS process in drugs industry based on over 600 tasks to be optimally scheduled. The conclusions in this first part of the study reveal the complexity of MOFJSSP and the analysis outcome of two optimization models for MOFJSSP: waiting systems and STRIPS language. The results of the research (part I and II) point out the limits or critical aspects of all the mentioned optimization models and the process characteristics that claim certain model as the most adequate.