Particle swarm optimization in multi-user orthogonal frequency-division multiplexing systems

• Main Author: Lye, Scott Carr Ken
• Format: Thesis
• Language: English; English
• Published: 2013
• Subjects: TK5101-6720 Telecommunication Including telegraphy, telephone, radio, radar, television
• Online Access: https://eprints.ums.edu.my/id/eprint/41471/1/24%20PAGES.pdf; https://eprints.ums.edu.my/id/eprint/41471/2/FULLTEXT.pdf; https://eprints.ums.edu.my/id/eprint/41471/
• Institution: Universiti Malaysia Sabah

Orthogonal Frequency Division Multiplexing (OFDM) is a multicarrier modulation technique of transferring an information bit stream on several sub-carrier frequencies. OFDM is used in many communication systems which exhibit high spectral efficiency and robustness against multipath fading channels. However, scheduling and resource allocation in multiuser OFDM system is complicated due to the different possibilities faced by the sub-channel assignments, furthermore the requirements for each user is not homogeneous. Therefore, it is challenging to strategically allocate radio resources and maximize system performance in a multiuser environment. Through modelling of the multiuser OFDM communication system, investigation in simulations shows that adaptive modulation in OFDM uses Channel State Information (CSI) to optimize the sub-carrier modulation scheme. While maintaining a target Bit Error Rate (BER), adaptive modulation optimizes the selection of modulation scheme and transmit power for each sub-carrier so that spectral efficiency is maximized. In this research, the overall system performance improvement is achieved by allocating the best user-to-sub-carrier combinations. To minimize the power consumption, Particle Swarm Optimization (PSO) is utilized to find the exact or near optimal resource allocation for the users. PSO is efficient in handling big solution space akin to resource allocation problems with different permutations mentioned. As a part of enhancing the performance of PSO, investigation of the control parameters effect on multi-user OFDM resource allocation is presented, resulting in particle reselection and dynamic inertia approach which shows 8 % of improvement over the standard PSO algorithm. Results also prove that the combination of both enhancements helped the algorithm to perform significantly better compared to a single enhancement. Furthermore, the introduction of PSO showed 70-75 % of power saving advantage over suboptimal resource allocation techniques.