Fuzzy Knowledge Optimization And Its Application Based On Improved Niching PSO Algorithm

Knowledge is an important resource in our day-to-day activities. It is continued growing and improving. Since then, knowledge optimization is highly necessary.Optimizing knowledge to provide more dependable and more complete knowledge is an effective means on improving process of decision-making and management.Algorithm for optimizing knowledge has always been one of the core parts of knowledge optimization. Among these, the swarm intelligence optimization algorithm has been highly favored. And on the other hand, the conception of fuzziness is increasingly introduced in various fields. For many productions now are difficult to be accurately described but with a more or less ambiguity. Thus, work of this paper mainly focuses on the optimizing algorithm of fuzzy knowledge.Firstly, the particle swarm optimization algorithm is introduced. The algorithm is applied to optimize fuzzy knowledge. Time-varying acceleration factor, circulation walls and other strategies are used to improve search performance. To improve the efficiency of the algorithm, some changes are made to the coding which brings much convenience. Experimental results show that this method can effectively optimize the fuzzy knowledge base.Since the particles inherent ‘homoplasy', swarm tends to gather together through the iterative process. This causes the loss of diversity. This paper introduces the fuzzy C-means clustering algorithm for sample optimization. Samples will be optimized while keeping the distribution. Meanwhile, according to cluster centers, discrete subsets are got since samples are divided. Niches are created then. This separates niching particle swarm optimize algorithm is then applied to optimize fuzzy knowledge base.It can be seen through experimental that the proposed method is an effective way for avoiding sticking at local optima.To combine theory with practice, this paper introduces programming learning styles diagnostics. The proposed optimization strategy is be applied to programminglearning styles diagnose problems. The presented method is applied to optimize fuzzy rule base to increase the performance of four-dimensional leaning style diagnosis.