||The last two decades have seen the advent of metaheuristic-based methods to solve both continuous and discrete optimization problems. While there are many paradigms available, the Particle Swarm Optimization (PSO) algorithm has had a profound impact in the mathematical optimization field. Extensive efforts, often with good results, have been made in the literature to improve the performance of the PSO algorithm by modifying the inertia weight, the position, and the velocity update equations. While these equations are still fundamental in its discrete counterpart, the Binary Particle Swarm Optimization (BPSO) algorithm, they are secondary to the velocity mapping procedure, which is essential in the discrete optimization process. |
Therefore, this thesis proposes a Modified Binary Particle Swarm Optimization algorithm (VS-BPSO) to solve the first stage of the forward day-ahead Thermal Unit Commitment Problem by considering two families of velocity mapping functions, namely the S-Shaped and V-shaped families. A novel pivot heuristic repair algorithm which effectively handles the conflict between system-wide and time-dependent constraints, is used to obtain a feasible Unit Commitment (UC) schedule. The Economic Dispatch (ED) subproblem is then solved for each time slot of the study period with an Enhanced Lambda Iteration Method (ELI).
The viability of the proposed solution methodology is simulated in the MATLAB © environment and is verified with the use of two benchmark systems, namely the IEEE 10-unit and the IEEE 26-unit test case systems. The obtained numerical results demonstrate the enhanced performance of the VS-BPSO algorithm in terms of total cost and computational time when compared to other methods in the recent literature. Furthermore, three additional cases involving other constraints and objectives are included. Comprehensive discussion of the obtained results is provided.