Optimization Algorithm Of Backbone Bacteria Foraging And Its Application In Portfolio Problem

Swarm intelligent bionic optimization algorithm is a kind of random search algorithm that simulates the evolution of natural organisms or social behavior of groups.It is often used to approximate some optimization problems that are difficult to solve directly,such as NP-hard problems.Now more and more scholars have optimized and improved swarm intelligence bionic algorithm,and achieved good results.Bacterial foraging optimization algorithm is a kind of intelligent bionic algorithm,which solves the complex nonlinear optimization problem by simulating and modeling bacterial foraging behavior.Bacterial foraging optimization algorithm is easy to handle and does not need to optimize the gradient information of the object.It has been widely used in automation engineering,production scheduling,image processing and other fields.This paper introduces the backbone particle swarm optimization algorithm and standard bacteria foraging optimization algorithm in detail.Inspired by backbone particle swarm optimization,an adaptive backbone bacteria foraging optimization algorithm(BBBFO)is proposed.This improvement makes full use of the strategy of randomly generating new solutions in the neighborhood of high-quality solutions of the backbone idea in the algorithm,and introduces the backbone operation based on dynamic Gaussian mutation into the breeding operation,which not only increases the population diversity on the basis of maintaining the elite neighborhood search,but also improves the solution accuracy.In addition,the fixed swimming step size in the chemotactic operation is adaptively improved,which makes the search range of the algorithm wide in the early stage and gradually narrow in the later stage,so that the algorithm pays more attention to the global and local search in different stages of solving the problem,and keeps the overall balance.In order to verify the adaptive optimization algorithm of bacteria foraging,the CEC2014 benchmark test function was used to carry out simulation experiments,and compared with other bacteria foraging algorithms,it was proved that the improved algorithm proposed in this paper had stronger global searching ability and overall performance.Based on an adaptive optimization algorithm of backbone bacteria foraging,three other backbone structures are proposed and summarized in detail in this chapter.In order to compare four optimization algorithms of backbone bacteria foraging online,23 CEC2014 benchmark test functions are selected for simulation experiments.The significance of portfolio investment lies in correctly selecting assets and investing among a large number of assets,and balancing the benefits and risks.This paper introduces the theory of portfolio optimization,and gives the measurement of returns and risks of single and multiple assets,Markowitz mean-variance model and mean-var model.By introducing risk aversion coefficient,the effective frontier of mean-variance model and mean-var model is obtained.In this paper,the mean-VaR model is used to apply the improved algorithm to the portfolio problem,and the comprehensive performance of the new algorithm proposed in this paper and the classical bacterial foraging optimization algorithm is further compared and verified.By randomly selecting the closing data of 50 stocks for one year,the simulation results show that BBBFO has better convergence speed and accuracy,as well as better stability and robustness than BFO.In addition,we also study the influence of different yield thresholds and confidence levels on VaR,and draw the fact that high-yield and high-risk investments are made.For prudent investors,we can adjust the yield thresholds to control investment risks so as not to fall into despair.For risk-averse investors,the portfolio with higher confidence level is preferred.