Single-objective and multi-objective optimization algorithms based on sperm fertilization procedure / Hisham Ahmad Theeb Shehadeh

• Main Author: Hisham Ahmad, Theeb Shehadeh
• Format: Thesis
• Published: 2018
• Subjects: QA75 Electronic computers. Computer science; QA76 Computer software
• Online Access: http://studentsrepo.um.edu.my/11831/1/hisyam_PHD.pdf; http://studentsrepo.um.edu.my/11831/2/hisham.pdf; http://studentsrepo.um.edu.my/11831/
• Institution: Universiti Malaya

In this work, Single Objective Optimization Algorithm (SOOA) is proposed. The SOOA version is extended to Multi Objective Optimization Algorithm (MOOA). To demonstrate the applicability of the proposed MOOA, a set of Wireless Sensor Network (WSN) problems is optimized. In SOOA, a novel metaheuristic approach based on a metaphor of a natural fertilization procedure, called “Sperm Swarm Optimization (SSO)” is proposed. In this approach, an optimization model of a sperm fertilization procedure is devised. The model follows the characteristics of sperm swarm, which moves forward from a low-temperature zone called Cervix. During this direction, sperm searches for a high-temperature zone called Fallopian Tubes where the egg is waiting for the swarm to fertilize at this zone, which this area is considered as the optimal solution. The SSO is tested with several benchmark functions used in the area of optimization. The obtained results are compared with the results of four algorithms. These algorithms are Genetic Algorithms (GA), Parallel Genetic Algorithm (PGA), Particle Swarm Optimization (PSO) and Accelerated Particle Swarm Optimization (APSO). The results show that the proposed SOOA outperformed other SOOAs algorithms in term of convergence and quality of the result. Then, the SSO has been extended to MOOA, called “Multi-Objective Optimization Algorithm Based on Sperm Fertilization Procedure (MOSFP)” depends on Pareto dominance, mutation operations and a crowding factor, that crowd and filter out the list of the best sperms (global best values). The proposed MOSFP is compared against three well-known MOOAs in the field of optimization. These algorithms are SPEA2, NSGA-II, and OMOPSO. The experimental results show that the efficiency and performance of the proposed MOSFP are highly competitive, which outperformed both of SPEA2 and NSGA-II algorithms in solving all the problems. In addition, the proposed MOSFP outperformed OMOPSO in solving problems such as WFG5, WFG8, and ZDT3. At the end, the proposed MOSFP has been used to solve a real-life problem such as optimizing a set of Quality of Services (QoS) objective functions (network models) in WSN. These objective functions are end-to-end latency, end-to-end delay, energy efficiency and network throughput. The optimal value of packet payload size that able to maximize the energy efficiency and network throughput as well as to minimize the end-to-end latency and end-to-end delay is sought. The result of the proposed MOSFP is compared against SPEA2, NSGA-II, and OMOPSO. Different packet payload sizes are supplied to the algorithms and their optimal value is derived. From the experiments, the intersection point and the knee point of all the obtained Pareto fronts for all the algorithms show that the optimal packet payload size that balances and manages the trade-offs between the four network models is equal to 45 bytes. The results also show that the performance of our proposed MOSFP is highly competitive and have the best average value compared to the other three algorithms. Furthermore, the overall performance of MOSFP from four models outperformed SPEA2, NSGA-II, and OMOPSO by 51%, 6% and 3% respectively.