Swarm Intelligence Algorithm Inspired By Twitter

In recent years,the researchers have proposed a series of swarm intelligence(SI)algo-rithms by mimicking the intelligent behavior of biological groups in nature.As heuris-tic algorithms,SI algorithms can complete the optimization task without calculating the derivative of the objective function or even without the formula of the objective function.Thus SI algorithms have been widely used in the optimization field because of it's universality compared with gradient-based algorithm.However,existing SI algorithms still suffer from local convergence and high time complexity.Therefore,a new sSI method named Twitter Optimization is proposed in this paper.As we all know,the information dissemination in Twitter network has an optimization tendency:the hottest information currently can be exploding in a short time through Retweeting behavior of each Twitter user,while the low-quality infor-mation is just the opposite.Therefore,we simulated the crowd behavior on Twitter to design two SI algorithms to solve the above problems in traditional algorithms.The main contribution of this article are as follows:?In order to solve the problem of falling into local optimum in traditional SI algo-rithms,we proposed Twitter Optimization(TO)algorithm.TO abstracts Tweet information as the solution vectors of the objective function.Tweeting,Retweet-ing,Following and other behaviors on Twitter are abstracted as computing s-trategy.Thus TO keeps the balance between exploration and exploitation and becomes hard to fall into local optimum.?In order to solve the problem of high time complexity in solving multi-objective optimization problems with traditional SI algorithms,we proposed multi-objective Twitter Optimization(MOTO)algorithm.Based on TO,MOTO introduces the characteristic of individual's Preference in Twitter and combines it with the weight-ed sum decomposition method.By making individuals with similar Preference follow each other,individuals with similar weights gradually connect to each other instead of computing k nearest neighbors.Thus MOTO reduces it's time complexity.?At last,we use the above two optimization algorithms to solve the path plan-ning problem in our simulation environment.By translating the path planning problem into optimization problem,we solved the single-objective path plan-ning problem by TO and the multi-objective path planning problem by MOTO and completed the visualization of this experiment in a self-built 3D simulation environment.TO and MOTO have improved the accuracy and stability of path planning compared to traditional SI algorithms through their excellent optimiza-tion capabilities.