Summary: | 博士 === 中原大學 === 電子工程研究所 === 100 === Heuristic algorithm is structured by natural phenomena with some rules and randomness.
It helps us to calculate the optimal solution of complex problems by computer. Heuristic
algorithm has been widely used in the search, optimization, scheduling and other engineering
problems. Heuristic algorithms can be divided into three parts, biological evolutionary
process, animal behavior and physical annealing process. Most heuristic algorithms focus on
how to search for optimal solution more effectively in the solution space, thus all algorithms
are faced with how to avoid falling into the local optima. Therefore, we attempt to propose
several strategies to enhance the performance of heuristic algorithms. The first, we proposed
the 2-Opt Algorithm to enhance the performance of Differential Evolution. Besides, we also
proposed restricted randomization, cooperation and cultural activation strategies to enhance
the performance of Particle Swarm Optimization. The experimental results show the
proposals outperform original DE and variant PSO in terms of solution accuracy and
convergence speed.
|