A combined multi-agent and case-based reasoning approach to support collaborative nonconformance problem solving in the thermoplastic injection moulding process

Abstract

Injection moulding is one of the most versatile technologies for processing thermoplastic materials used nowadays in industry. Despite all the technological resources used, the probability of occurring nonconformances in the production stage cannot be completely eliminated, and in general these nonconformances occur owing to the complex interrelation among the constructive elements of the moulded part, the mould, the rheological characteristics of the resin, and the processing conditions. These factors lead to a difficult analytical modelling, and frequently cause the task of discovering the source of nonconformance and its solution to become slow and costly. Owing to this scenario, this paper presents an agent-based architecture that includes case-based reasoning agents to support nonconformance problem solving. This paper presents the set of descriptors that represent the nonconformance features, as well as the implementation of a prototype based on the Java Agent DEvelopment Framework (JADE Platform) and the jCOLIBRI CBR Framework as core engine. The results show that the developed system supports adequately the collaborative nonconformance problem solving in injection moulding, since it quickly provides, through a friendly interface, the following information: (a) knowledge inference, retrieval and presentation; (b) the suggested solutions for the nonconformance; (c) the result of carrying out the solution.