Rearch On Continues Optimization Problem Fitness Landscape Analysis And Application Base On PSO

Continuous optimization problem belongs to a large class of optimization problems, and the use of heuristics to solve them, is currently a widely used means. Operating efficiency is determined jointly by the heuristic algorithm design and structural problems, however, the lack of an effective structure for problem analysis methods to make continuous optimization problem solution space the lack of appropriate information and knowledge, leading to the application of algorithms and problem not necessarily match, resulting in low efficiency solving.Fitness geomorphological analysis incorporated herein attempts to widely used combinatorial optimization problems in continuous optimization problems, the paper proposes a method for continuous particle swarm algorithm optimization analysis of fitness landscapes, through fitness geomorphological features of the problem analysis, and This is based on a targeted selection and design algorithms for problem solving. First, well-proven algorithms on the test function does not necessarily apply to new problems, new problems that may not be able to enjoy a free lunch old algorithm. Secondly, the proposed combination of probes PSO fitness landscapes depicting method, PSO run in the solution space when multiple individuals stay in the same position as the search for optimal solutions based on local extreme draw adapt the solution space degree landscape. Finally, this approach will be used in the artificial neural network weights training, proven solution space of artificial neural networks fitness landscapes have stepped decline characteristics and features designed for this algorithm targeted, prove that algorithm design more effective than universal improvement in handling weights artificial neural network training.