A hybrid GA –SA multiobjective optimization and simulation for RFID network planning problem

Abstract

Optimized asset tracking with Radio Frequency Identification (RFID) as a complicated innovation that requires much money to be implemented has become more popular in the healthcare industry. Considering the use of more antennas in each reader, we present a modern heuristic methods, hybrid of Genetic Algorithms (GA) and Simulated Annealing (SA) for the purpose of placing readers in an emergency department of a hospital with an RFID network. In this study, a multi-objective function is developed for the network coverage maximization and the minimization of total cost, tag reader collision, interference, energy consumption, and path loss in a simultaneous way. The proposed algorithm provides savings (on average) in the total cost of the RFID network through the efficient use of three types of readers with one, two and four antenna ports. Additionally, by testing three scenarios, the effect of algorithms in achieving the optimal solution is indicated by the simulated results. Besides, the results of GA-SA is compared to the results of GA and other existing models in the relevant literature. It is shown that its main advantage is the use of multi-antenna RFID readers, which reduces the total cost of the RFID network and also increase network coverage with fewer readers and antennas. In other words, contributions for the research are proposing a hybrid GA-SA algorithm, developing a multi-objective function, testing the algorithm in a hospital setting, and comparing the results of GA-SA with GA.

Keywords

• genetic algorithm (ga)
• simulated annealing (sa)
• asset tracking
• multi-antenna readers
• hospital