Optimal RFID networks scheduling using genetic algorithm and swarm intelligence

Abstract

RFID is an emerging technique for identifying items and all kinds of real world applications. Multi RFID readers are implemented to the product line in many industries and they consist of varied reader resources. But there are some defects with the disposition of the RFID-based application. The phenomenon of incorrect negative reads occurs in a multi-tag and multi-reader environment where a tag that is present is not detected. Collisions occurring between readers cause the faulty or missing reads. The stopgap is to solve the frequency allocation problem for networks of RFID readers. Furthermore, finding the optimal structure of readers and scheduling the readers to reduce the total system transaction time or response time are both challenging problems. In the presence of interdependencies, the optimal scheduling problem to minimize the overall transaction or response time is modeled as a graph partitioning problem (GPP). GPP is a well known NP-complete problem. The more readers exist in the product line, the higher complexity of the problem. Designing a schedule having the maximum parallelism reduces the total transaction time but may not minimize it. In this research, we integrate genetic algorithms with binary particle swarm optimization (GA-BPSO) to solve the Multi RFID networks scheduling problem. Simulation results on a real-world problem show that the GA-BPSO algorithm provides robust solution quality and is suitable for scheduling large scale RFID reader networks.