VOL. 12, NO. 3, FEBRUARY 2017 ISSN 1819-6608 ARPN Journal of Engineering and Applied Sciences ©2006-2017 Asian Research Publishing Network (ARPN). All rights reserved. www.arpnjournals.com 834 HYBRIDIZED FIREFLY ALGORITHM FOR MULTI-OBJECTIVE RADIO FREQUENCY IDENTIFICATION (RFID) NETWORK PLANNING Adel Muhsin Elewe, Khalid Bin Hasnan and Azli Bin Nawawi Faculty of Mechanical and Manufacturing Engineering, UniversitiTun Hussein Onn Malaysia (UTHM), Parit Raja, BatuPahat, Johor, Malaysia adelmuhsin2@gmail.com ABSTRACT The RFID network planning (RNP) problem belongs to the large-scale multi-objective hard optimization problems. RNP aims to optimize the overall read region based on a set of objectives. A novel approach of hybrid firefly algorithm was developed for multi-objective RNP problem. The technique was combining the Density Based Clustering method (DBSCAN) and firefly algorithm. Empirical tests were conducted on six standard RFID benchmark sets with random and clustered topologies. A comparative analysis performed with other state-of-the-art algorithms based on the same test data.Simulation results exhibited uniformly better performance in achieving maximum coverage with smaller number of deployed readers and less transmitted power. Keywords: RFID networks planning, firefly algorithm, BDSCAN. INTRODUCTION Development in manufacturing technology has enabled radio frequency identification (RFID) applications in various fields such as automatic detection and supply chain management. RFID system transmits data by small electronic chip, called a tag, while the RFID reader collects the data from tagged items and transmits to the middleware in electromagnetic wave environment (Hasnanet al. 2015). The central optimal RFID network parameters include tag coverage, interference, economic efficiency and load balance. All these parameters are related to two origins, the topological tag distribution and the limited range of RFID reader (Hasnanet al. 2015) (Lu and Yu 2014). Several algorithms have been developed to optimize RFID network planning. Particle Swarm Optimization (PSO) algorithm was the well-known optimization technique due to fast operation speeds, ease of implementation and fewer parameters that require adjustment (Elewe et al. 2016). The latest development for PSO algorithm was Multi Colony – Global Particle Swarm Optimization (MC-GPSO) (Hasnanet al. 2015). The state of art optimization-based approach includes many other recently proposed methods such as firefly algorithm (Bacanin/a et al. 2015), ABC algorithm (Bacanin/b et al. 2015), and fireworks algorithm (Tuba et al. 2015). The simulation results of these methods demonstrate the more efficient optimal solution of RFID Network Planning (RNP). This paper presents the development of clustering strategy based on RFID circular reader range to specify the number readers that can cover all the tags distributed at two-dimensional planes as a pre-process operation in order to improve the RFID network planning performance. The proposed method discretized the distributed tag area based on the ability of reader coverage and clusters the tag into groups based on the Density-Based Algorithm. The suggested model in this method used the DBSCAN algorithm (Density Based Spatial Clustering of Applications with Noise) which is designed to discover the clusters in a spatial database (Bäcklundet al. 2011). This method was used to specify the appropriate number of required readers and enhanced the operation of the system. The DBSCAN method is combined with firefly algorithm (FA) as a pre-process step to solve the RNP issues by deploying minimum number of readers to cover all tags with minimum interference between readers and requires less transmitted power. MATHEMATICAL FORMULATION This section presents the mathematical definition of the RFID network-planning problem that was also used recently by Hasnan in 2015. The RFID system is a Radio communication between tags and readers and involves writing/reading information between them. To deploy RFID network, three important questions must be answered (Chen et al. 2011): how many readers are needed, what is the optimal location for the readers, and how should readers’ parameters be adjusted. The answer to these questions can be detected from the set of equations that are concerned with the objective functions of RNP problems. The first and one of the most important objectives employed in this model is optimal tag coverage (C) which enables the ability to detect and obtain the IDs of all of the deployed tags (Botero and Chaouchi 2011). It can be considered the sum of the difference between the actual power received by each tag to the required power and is formulated as: ) 1 i ( min        req T tagi C (1) Ptagi = Actual received power at each tag Preq = required threshold power NT =Number of tags in working area The Friis transmission equation power at each tag can be calculated by the following equation (Hasnan et al. 2015):