Heuristic Optimization Algorithms

Project Title:

Heuristic optimization algorithms with application to power system

Project Description:

Optimization problems are widely encountered in various fields. Many traditional optimization methods, such as quadratic programming and nonlinear programming, have been used to solve these optimization problems. However these optimization methods above have disadvantages to some extent such as getting trapped in local optima, increased computational complexity, and its inability to solve multi-objective optimization problems. Heuristic optimization algorithms are fast growing methods that can overcome the restrictions as imposed on the traditional methods, and have received much attention regarding their potential as global optimization algorithms.
Increasing demands for energy, growing concerns about energy sustainability and environmental protection, coupled with market and regulatory pressures, all contribute to a more stringent requirement for economic operation of power systems, where economic dispatch has a significant role to play. Economic dispatch involves minimizing the cost while meeting the energy requirements over an appropriate period and at an acceptable level of reliability. Conventional economic dispatch of power system involves only thermal energy generation. The objective of economic dispatch is to minimize the total cost of supplying required energy while distributing the load demand among on-line generators. Due to physical characteristic limits of generators and power systems, economic dispatch is subjected to certain constraints involving transmission losses, valve-point effects for multi-valve generators, ramp rates or prohibited operating zones etc. Economic dispatch incorporates two steps: one is pre-dispatch or selection of generators to meet the power load; another is economic dispatch determining the power output of generators on. A range of traditional methods have been employed to solve practical economic dispatch problems, such as dynamic programming, quadratic programming and nonlinear programming. More recently, evolutionary algorithms, such as genetic algorithms, evolutionary programming, tabu search, differential evolution and particle swarm optimization (PSO) have also been studied due to their potential for global optimization.
Currently, all proposed heuristic algorithms, such as particle swarm optimization, differential evolution and genetic algorithm, have several parameters to be adjusted simultaneously when employed. Given a set of design parameters, these algorithms may work well on one set of problems but may perform badly on another or, even for the same set of problems the results can be quite different using different constraints handing technologies or as the dimension grows. My research aims to further improve the existing heuristic algorithms such as PSO and DE, propose new algorithms, and apply these techniques to electric power system.