Research On Environmental Detection And Response Strategy For Dynamic Optimization Problems

Dynamic optimization is an important method to achieve energy conservation,consumption reduction,as well as value gain,and therefore it has been widely used in many fields such as transportation,industry,economy and life.The environment of dynamic optimization problems is constantly changing,consequently,the detection and response to environment plays a key role in improving the adaptability of optimization algorithms and also has become one of the main research directions of dynamic optimization problems.Moreover,dynamic optimization problems are easily affected by random factors in practice,so it has important theoretical significance and practical value to consider those kind of dynamic optimization problems possessing randomness.This thesis proposes an adaptive environmental detection and response strategy to overcome the shortcomings of existing methods,such as low accuracy of detection methods and blindness of response strategies.In addition,there isn’t much research work on stochastic dynamic optimization problems so far,this thesis proposes a new type of stochastic dynamic optimization problems,which optimal solutions are subject to random disturbances.By applying adaptive environmental detection and response strategy put forward,stochastic dynamic optimization problems can be solved effectively.The main research work is as follows:1.The orthogonal table interval shrinking method is improved.This thesis improves the existing orthogonal table interval shrinking method from the selection of test points and the interval shrinking method.On this basis,the function optimal solution interval shrinking method is constructed and used to determine the region where the optimal solution of the multivariate function is located.Then simulation experiments verify the effectiveness of this method for unimodal and multimodal functions,and the comparison experiment with existing method also show that this method has higher accuracy.2.Based on the function optimal solution interval shrinking method,we propose an adaptive environmental detection and response strategy for dynamic optimization problems.In terms of environmental detection,this thesis selects the exploration particles in a certain proportion inside and outside the optimal solution areas of the fitness value candidate functions,and detects the environmental changes according to the changes of the fitness value of the exploration particles.In terms of environmental response,a memory mechanism is introduced to store environmental information.After the environment changes,according to the environmental information and the change of the optimal solution,particles are selected in the changed optimal solution area to adapt to the new environment.The experiment results prove that the environment detection and response strategy improves the accuracy of the environmental detection and the speed of the algorithm to converge to the optimal solution after the environment changes.3.A new type of stochastic dynamic optimization problems is proposed,which optimal solutions are subject to random disturbances.And the environmental detection and response strategy in this thesis is applied to the solution of such stochastic dynamic optimization problems.Under the assumption that the random disturbances obey the normal distribution,we give the upper limit of the standard deviation of the random disturbances corresponding to no change in the environmental detection before and after the disturbance.Then experiments verify the effectiveness of the proposed environmental detection and response strategy in dealing with such stochastic dynamic optimization problems.