Grey Wolf Optimizer And Its Applications

Optimization problems exist widely and Grey Wolf Optimizer(GWO)provides complex optimization problems with a new idea and method.GWO is a new swarm intelligence technique mimicked the leadership hierarchy and hunting mechanism of grey wolves in nature.GWO has simple structure and clear concept,moreover,it is easy to implement and has good global performance.However,the research and applications of GWO are still at its primary stage.There are still many key problems to be solved,mainly including slow convergence rate at later stage and weak local capability.In this dissertation,the GWO is researched deeply where some shortcomings existing in GWO are analyzed and improved and some new and better performance methods are proposed.The aim is to perfect and broaden the theorical basis and application range of GWO,which provides an effective new method for solving large-scale optimization of complex systems.The main contributions of this dissertation are as follows:(1)This dissertation analyzes the shortcomings of the mathematic theory base and application of GWO,and puts forward some improved strategies and methods.GWO is firstly used to solve the problem of unmanned combat aerial vehicle path planning(UCAV).Compared with the current CS,FPA,NBA,BSA,ABC,and GGSA methods,the results show that GWO can obtain higher quality solutions which have certain direct significance and referential value for designing UCAV.(2)In order to better balance the global and local search capabilities of GWO,orthogonal design strategy and differential mutation strategy are introduced.An orthogonal grey wolf optimizer with mutation(OMGWO)is proposed.The orthogonal design operator is introduced before the grey wolf individual updates its location.The global search ability of GWO is enhanced by orthogonal design operator.At the same time,the differential mutation operator "DE / best / 2 / bin" is introduced in later iterations of GWO,which improves the local search ability of the algorithm.OMGWO can effectively optimize the neural network structure,performance problems by verifying five standard datasets.(3)In order to improve the convergence rate of GWO,the basic principle of ranking-based mutation is studied.The ranking-based mutation strategy is introduced before the grey wolf individual location update.Grey wolf optimizer with ranking-based mutation operator(RGWO)is proposed.Moreover,three IIR instances are used to verify the feasibility and validity of RGWO solving model identification problem.(4)In order to break through the traditional thinking of selecting image preprocessing according to different background conditions,the mathematical model of lateral inhibition is studied.The lateral inhibition in vision is applied to image preprocessing to achieve the purpose of adaptive preprocessing.GWO is used to perform the searching task on the feasible domain space of the image preprocessing.A grey wolf optimizer with lateral inhibition(LI-GWO),which is applied to template matching problem,is proposed.The experiment shows that LI-GWO can effectively solve template matching problem and has fast convergence rate and high accuracy.