Research On Continuous And Discrete Application Problems Based On Heuristic Algorithm

Heuristic algorithms are intelligent algorithms designed to imitate the behavior of biological populations,inspired by the thinking of organisms,including humans,in handling affairs.The traditional exact solution algorithm is often difficult to solve when faced with some complex NP problems.The heuristic algorithm can give a feasible solution to the optimization problem with limited time and space cost.The lion swarm optimization algorithm is a heuristic algorithm designed and invented by researchers in recent years to imitate the behavior of lion colonies.Lions with different identities have different behaviors.The lion king leads the way,the lioness cooperates in hunting,and the cub eat near the lion king,follows the lioness to learn to hunt,or be driven out of the territory to become a stray lion.Compared with some classical heuristic algorithms,the lion swarm optimization algorithm includes information exchange between different sub-populations,and the methods are more diverse and the mechanism is more flexible.According to the characteristics of continuous and discrete application problems,this thesis improves the lion swarm optimization algorithm and solves the application problems.For continuous optimization application problems,the definition domain of the independent variable of the optimized individual is the continubus real number domain.Aiming at the problem that the position update formula of the lion group algorithm is sensitive to the symmetric domain of the origin and easy to fall into the local extreme value,an improved lion swarm optimization algorithm is proposed.Combined with the migration and spiral search behavior of the seagull algorithm,the collision of individual positions is avoided,and the search accuracy of the lion swarm optimization algorithm is improved;combined with the supervision mechanism,the algorithm can be prevented from falling into local extreme values for a long time and wasting computing power.The improved lion swarm optimization algorithm was tested on the international standard test function set CEC2017,and the optimization effect was better than the comparison heuristic algorithms.For the housing price prediction problem,the BP neural network method is used to solve it.Aiming at the shortcomings of BP neural network being sensitive to initial value and easy to fall into local extreme value,the lion swarm optimization algorithm is used to optimize the weight and bias of BP neural network.The experiment uses the Qingdao Jimo housing price data set captured on the internet for training and testing,and the experiment proves that the model works well.For discrete optimization application problems,the encoding method of the optimization individuals is a discrete integer sequence.The original lion swarm optimization algorithm’s independent variable domain is a continuous real number interval,which is not applicable in this case.The vehicle routing planning problem optimizes a sequence of integers,and the method of simply converting the optimization variables into binary numbers is not applicable.Therefore,the basic framework of the lion swarm optimization algorithm is retained,and a discrete lion swarm algorithm is designed by combining the multi-neighborhood structure crossover mechanism and the diversity measurement mechanism.For the vehicle routing problem with dual constraints of loading and capacity,the constraints increase,which requires the algorithm to have strong global search ability and the ability to jump out of local extreme values.The improved lion swarm optimization algorithm is combined with the least open space method,and the strategy of verifying first and then loading is adopted to avoid repeatedly calling the packing module and reduce the efficiency of the algorithm.The experiment is tested on the international classic data set,the result show that algorithm has good effect.