The Development And Application Of Particle Swarm Optimization (PSO)

Many chemical engineering problems involve the optimization of a set of parameters with the aim of minimum or maximum the objective function. Furthermore, it is required many optimization approaches in the process of modeling.Traditionally, solutions to optimization problems are often using either problem-specific heuristics or variants of the local-search method that iteratively refines a single candidate solution to the problem. Unfortunately, the traditional methods are not robust with respect to problem-type and often only work on well-defined problems where the number of possible solutions is not too large. As a consequence, they often fail either in terms of computing-time or quality of the found solution on interesting real-world problems that usually have a large search space, and neither is well-defined.Population-based optimization algorithms have shown capabilities of approximating optimal solutions to these real-word problems within a reasonable amount of time. The best known of these algorithms are the genetic algorithm (GA) and ant colony optimization (ACO) algorithm. GA is inspired by natural evolution, and the inspiring source of ACO is the best pheromone trial laying and following behavior of real ants which use pheromones as a communication medium. GA is one of the evolutionary algorithms, while ACO belongs to the swarm intelligence algorithms.Particle swarm optimization (PSO) algorithm was developed by Kennedy and Eberhart in 1995, which has obvious ties with both evolutionary computation and swarm intelligence. The underlying motivation for the development of PSO algorithm was social behavior of animals such as bird flocking, fish schooling, and swarm theory. PSO is similar to GA in that the system is initialized with a population of random solutions, and the potential solutions, called particles, are then "flown" through theproblem space. Each particle keeps track of its coordinates in the problem space which are associated with the best solutions it has achieved so far and obtained so far by any particles in the population. The attractive character of PSO concept is quite simple and easy to implement. Although, at the time of writing this dissertation, PSO has been developed around for ten yeas, many aspects required to be further investigated whatever theory or practice.In this work, an hybrid particle swarm optimization (HPSO) was proposed, where the Hooke-Jeeves pattern search is combined to PSO to speed up the local search, also mutation operation is embedded to avoid the common defect of premature convergence. Two thresholds were adopted to balance the exploration and exploitation abilities. The performance of new algorithm was demonstrated through extensive benchmark functions and compared with that by the PSO. The obtained results showed the local search ability is improved, and the probability of finding the global optimal value by HPSO is larger than that by using PSO. HPSO is applied into FCC, and the result optimized by HPSO is better than that by SPSO. In the end, a method which employs PSO to solve constrained optimization problems was proposed.