Multi-Objective Optimization Approaches for Heterogeneous Systems

• Other Authors: Thevarajan, Thabendra (authoraut)
• Format: Others
• Language: English; English
• Published: Florida State University
• Subjects: Electrical engineering; Computer engineering
• Online Access: http://purl.flvc.org/fsu/fd/FSU_migr_etd-1592

The main goal of this master's thesis is to explore and develop centralized and optimization techniques for terrestrial power systems and for heterogeneous systems like a ship's cooling system. First, an agent based decentralized multi-objective optimization (DMOO) algorithm is presented. This algorithm use evolutionary programming as the fundamental optimization technique. The algorithm was implemented on three IEEE test systems (IEEE 14, IEEE 30 and IEEE 57 bus systems) and results are provided. The developed algorithm's performance was compared against a centralized multi-objective optimization (CMOO) algorithm, which also was developed as a bench mark for the performance of the DMOO. The CMOO algorithm was compared with earlier similar work found in literature. In addition, a CMOO algorithm was also developed for heterogeneous systems such as a ship's chill water cooling system. A small scale system representing a ship's chill water cooling system was designed and built for testing and validating the CMOO; The Table Top (TT) and TT Emulator. The TT used actual components such as valves, pumps and pipes. The Emulator is designed in Matlab/ Simulink that mimics the hardware based TT. The optimization algorithm was implemented on these systems for different scenarios and subsequent results were analyzed. The CMOO algorithm for power system has shown better results than the results it was compared with. The performance of the developed DMOO algorithm was equal that of CMOO algorithm for small systems and comparatively slightly lower than the CMOO algorithm's performance. The results of the CMOO algorithm for heterogeneous have proven the algorithm's ability in optimizing multiple objectives across various subsystems for a dynamically changing system in real time or near real time. === A Thesis Submitted to the Department of Electrical and Computer Engineering in Partial Fulfillment of the Requirements for the Degree of Master of Science. === Summer Semester, 2011. === June 21, 2011. === Multi-Objective Optimization, Decentralized Optimization, Heterogeneous Systems, Ship System, Power System Optimization === Includes bibliographical references. === Chris S. Edrington, Professor Directing Thesis; David A. Cartes, Committee Member; Sanjeev K. Srivastava, Committee Member; Rodney Roberts, Committee Member.