Improved Quantum Particle Swarm Optimization And Applied Research

Quantum-behaved Particle Swarm Optimization(QPSO) is a new model of the PSO algorithm. It is to be presented from the perspective of quantum mechanics. QPSO algorithm is a highly versatile optimization techniques can be applied to solve a variety of complex optimization problems. This paper analyzes the basic principle and algorithm model of PSO and QPSO algorithm, and proposed to improve the method of using chaotic optimization strategy and the normal cloud model, then the improved algorithm is applied to neural networks and prediction of oilfield development index.First, expounds the basic principle of traditional PSO algorithm, the existing problems and improvement measures. In Delta trap for example, analysis of the method of establishing QPSO iterative equations, compared with the traditional PSO, and the unique advantages of QPSO is presented.Second, the paper studies the QPSO existing problems and the ways to improve.Proposed particle swarm initialization method of chaos theory, use ergodic theory of chaos,this method allows the particles found in search for solution of the problem space. These improvements will be integrated application to QPSO improved, the effective enhance the performance of the original algorithm. As an example of standard function extreme optimization, the improved algorithm and the other four similar algorithm detailed comparison, the results demonstrate the effectiveness of the improved method.Third, the paper studied the engineering application of improved QPSO algorithm. First,the integration of improved QPSO and neural networks, as an example of the sunspot number forecast, study improved algorithm in a complex time series forecasting. Then, the detailed description of the oilfield development index prediction of related concepts,determination method, and sample data collection methods, research and the establishment of oilfield development index prediction model based on improved QPSO and BP neural network, and the predicted results were compared and analyzed. The results show that the method is effective and feasible.