An evolutionary algorithm for optimization of XML publish/subscribe middleware in electronics production

Computer aided manufacturing using XML (CAMX) production systems are built on Message-Oriented Middleware Frameworks, offering standards-based communication among machines and control software applications. CAMX Frameworks implement Publish/Subscribe of XML messages through an entity called the Message Broker (MSB). In order to create scalable frameworks, distributed MSB systems need to be deployed. In the distributed case, the topology optimization problem arises, as clients need to be assigned to one of several available MSB nodes. The problem is a variation of the graph partitioning problem, known to be NP-complete. The problem is augmented with weighted directed edges, multiple levels of cluster capacity, complex weight functions, and multiple optimization criteria. A solution was developed based on an evolutionary multi-objective algorithm. The developed algorithm was tested for optimizing two flexible electronics assembly systems used as case scenarios. Robust convergence time was bounded to a few seconds, demonstrating the suitability of the proposed algorithm for rapid topology reconfiguration in response to changes in the system