Extraction of association rules in the schematic design of product service system based on Pareto-MODGDFA

Product service system (PSS) forces the manufacturing enterprise to shift from simple product design to an overall solution design for products and services. Schematic design involves the combination of product modules and service modules. The existing rule mining methods often experience difficulties when extracting association configuration rules in the high-dimensional space of the product service system configuration. In this paper, we propose a new method to mine association rules by integrating a multi-dimension and multi-objective double group discrete Firefly Algorithm (MODGDFA) with Pareto rules. The association rule of schematic design is mainly extracted from a large number of historical design databases by using MODGDFA. The MODGDFA uses fitness function, support degree, confidence degree, interest degree, and comprehensibility to evaluate the quality of association rules. The Pareto-MODGDFA model is constructed to establish a relevant model of the PSS. In this method, discrete Firefly Algorithm coding is adopted. Two subgroups are updated according to their respective position updating method. The optimal fireflies of the two sub-groups exchange information to obtain different sets of non-dominated solutions and to extract the potential association rules from the database. The Pareto principle in the evolution process is used to evaluate the merits and demerits of the firefly in the multi objective functions and reduce redundant rules. The effectiveness and feasibility of the proposed method are demonstrated by using CNC machine tools of the product service system schematic design.