The system will be going down for regular maintenance. Please save your work and logout.
Stochastic Prediction with Particle Swarm Optimization
Murugesan, Ranjith Kumar
MetadataShow full item record
Scientific research involves random processes which are complex and unpredictable in nature. Higher the understanding of the randomness in the research field the better understanding of the process and more accurate the results. Optical electronics has its fair share of random processes that remain unexplained till date. The stochastic nature in the process such as spontaneous emission, polarization effects, Phase noises, non-linearity in losses are still unexplained. These processes contribute to the degradation of optical devices which could be avoided with a better knowledge of the unpredictable nature involved in these processes. This research is an attempt to study the interaction of random parameters incorporated within Particle Swarm Optimization (PSO) algorithm by using it as a standalone algorithm for prediction and estimation. PSO has been used in almost all possible fields of research due to its versatility, superior accuracy over other optimization algorithms and high convergence ratio with optimal initiation parameters. Through this research a new multidimensional model of PSO is proposed. This model uses the random factors involved in the environment as a parameter in the algorithm. PSO has been combined with factors influencing randomness in the environment to analyze, predict the possible outcomes in the future. The model has been tested for accuracy, adaptability and consistency. The foreign exchange market is used as a test environment. Possible applications and future work are also discussed.