Optimization of Maximum Entropy Model Oriented to Bayes Prior Distribution Based on PSO

2010 ◽  
Vol 37-38 ◽  
pp. 814-818
Author(s):  
Yi Dai ◽  
Zhong Min Wang ◽  
Guang Ling

How to acquire prior distribution is a key to Bayes method. Firstly, a nonlinear constrained optimal model of probability density function based on the principle of maximum entropy is set up. By using Lagrange multiplier this constrained optimal problem is transformed to a non-constrained optimal one, which is solved by standard particle swarm optimization (PSO) algorithm. Secondly, a new improved particle swarm optimization (IPSO) algorithm is proposed because standard PSO is slow on convergence and easy to be trapped in local optimum. IPSO introduces the hybrid method from genetic algorithm (GA) so that the overall searching ability is enhanced; then linear decreasing inertia weight is used to optimize particles. The simulation examples show that IPSO is simple and effective and it can rapidly converge with high quality solutions.

2012 ◽  
Vol 253-255 ◽  
pp. 1369-1373
Author(s):  
Tie Jun Wang ◽  
Kai Jun Wu

Multi-depots vehicle routing problem (MDVRP) is a kind of NP combination problem which possesses important practical value. In order to overcome PSO’s premature convergence and slow astringe, a Cloud Adaptive Particle Swarm Optimization(CAPSO) is put forward, it uses the randomicity and stable tendentiousness characteristics of cloud model, adopts different inertia weight generating methods in different groups, the searching ability of the algorithm in local and overall situation is balanced effectively. In this paper, the algorithm is used to solve MDVRP, a kind of new particles coding method is constructed and the solution algorithm is developed. The simulation results of example indicate that the algorithm has more search speed and stronger optimization ability than GA and the PSO algorithm.


2019 ◽  
Vol 18 (03) ◽  
pp. 833-866 ◽  
Author(s):  
Mi Li ◽  
Huan Chen ◽  
Xiaodong Wang ◽  
Ning Zhong ◽  
Shengfu Lu

The particle swarm optimization (PSO) algorithm is simple to implement and converges quickly, but it easily falls into a local optimum; on the one hand, it lacks the ability to balance global exploration and local exploitation of the population, and on the other hand, the population lacks diversity. To solve these problems, this paper proposes an improved adaptive inertia weight particle swarm optimization (AIWPSO) algorithm. The AIWPSO algorithm includes two strategies: (1) An inertia weight adjustment method based on the optimal fitness value of individual particles is proposed, so that different particles have different inertia weights. This method increases the diversity of inertia weights and is conducive to balancing the capabilities of global exploration and local exploitation. (2) A mutation threshold is used to determine which particles need to be mutated. This method compensates for the inaccuracy of random mutation, effectively increasing the diversity of the population. To evaluate the performance of the proposed AIWPSO algorithm, benchmark functions are used for testing. The results show that AIWPSO achieves satisfactory results compared with those of other PSO algorithms. This outcome shows that the AIWPSO algorithm is conducive to balancing the abilities of the global exploration and local exploitation of the population, while increasing the diversity of the population, thereby significantly improving the optimization ability of the PSO algorithm.


2019 ◽  
Vol 2019 ◽  
pp. 1-22 ◽  
Author(s):  
Hao Li ◽  
Hongbin Jin ◽  
Hanzhong Wang ◽  
Yanyan Ma

For the first time , the Holonic Particle Swarm Optimization (HPSO ) algorithm applies multiagent theory about the improvement in the PSO algorithm and achieved good results. In order to further improve the performance of the algorithm, this paper proposes an improved Adaptive Holonic Particle Swarm Optimization (AHPSO) algorithm. Firstly, a brief review of the HPSO algorithm is carried out, and the HPSO algorithm can be further studied in three aspects: grouping strategy, iteration number setting, and state switching discrimination. The HPSO algorithm uses an approximately uniform grouping strategy that is the simplest but does not consider the connections between particles. And if the particles with larger or smaller differences are grouped together in different search stages, the search efficiency will be improved. Therefore, this paper proposes a grouping strategy based on information entropy and system clustering and combines two grouping strategies with corresponding search methods. The performance of the HPSO algorithm depends on the setting of the number of iterations. If it is too small, it is difficult to search for the optimal and it wastes so many computing resources. Therefore, this paper constructs an adaptive termination condition that causes the particles to terminate spontaneously after convergence. The HPSO algorithm only performs a conversion from extensive search to exact search and still has the potential to fall into local optimum. This paper proposes a state switching condition to improve the probability that the algorithm jumps out of the local optimum. Finally, AHPSO and HPSO are compared by using 22 groups of standard test functions. AHPSO is faster in convergence than HPSO, and the number of iterations of AHPSO convergence is employed in HPSO. At this point, there exists a large gap between HPSO and the optimal solution, i.e., AHPSO can have better algorithm efficiency without setting the number of iterations.


Author(s):  
Rongrong Li ◽  
Linrun Qiu ◽  
Dongbo Zhang

In this article, a hierarchical cooperative algorithm based on the genetic algorithm and the particle swarm optimization is proposed that the paper should utilize the global searching ability of genetic algorithm and the fast convergence speed of particle swarm optimization. The proposed algorithm starts from Individual organizational structure of subgroups and takes full advantage of the merits of the particle swarm optimization algorithm and the genetic algorithm (HCGA-PSO). The algorithm uses a layered structure with two layers. The bottom layer is composed of a series of genetic algorithm by subgroup that contributes to the global searching ability of the algorithm. The upper layer is an elite group consisting of the best individuals of each subgroup and the particle swarm algorithm is used to perform precise local search. The experimental results demonstrate that the HCGA-PSO algorithm has better convergence and stronger continuous search capability, which makes it suitable for solving complex optimization problems.


2010 ◽  
Vol 20-23 ◽  
pp. 1280-1285
Author(s):  
Jian Xiang Wei ◽  
Yue Hong Sun

The particle swarm optimization (PSO) algorithm is a new population search strategy, which has exhibited good performance through well-known numerical test problems. However, it is easy to trap into local optimum because the population diversity becomes worse during the evolution. In order to overcome the shortcoming of the PSO, this paper proposes an improved PSO based on the symmetry distribution of the particle space position. From the research of particle movement in high dimensional space, we can see: the more symmetric of the particle distribution, the bigger probability can the algorithm be during converging to the global optimization solution. A novel population diversity function is put forward and an adjustment algorithm is put into the basic PSO. The steps of the proposed algorithm are given in detail. With two typical benchmark functions, the experimental results show the improved PSO has better convergence precision than the basic PSO.


2013 ◽  
Vol 427-429 ◽  
pp. 1710-1713
Author(s):  
Xiang Tian ◽  
Yue Lin Gao

This paper introduces the principles and characteristics of Particle Swarm Optimization algorithm, and aims at the shortcoming of PSO algorithm, which is easily plunging into the local minimum, then we proposes a new improved adaptive hybrid particle swarm optimization algorithm. It adopts dynamically changing inertia weight and variable learning factors, which is based on the mechanism of natural selection. The numerical results of classical functions illustrate that this hybrid algorithm improves global searching ability and the success rate.


2011 ◽  
Vol 130-134 ◽  
pp. 3467-3471 ◽  
Author(s):  
Bin Jiao ◽  
Zhi Xiang Xu

This paper proposes an improved particle swarm optimization algorithm (PSO) for the global and local equilibrium problem of searching ability. It improves the iterative way of inertia weight in PSO, using non-linear decreasing algorithm to balance, then PSO combines with simulated annealing (SA). Finally, the optimization test experiments are carried out for the typical functions with the algorithm (ULWPSO-SA), and compare with the basic PSO algorithm. Simulation experiments show that local search ability of algorithm, convergence speed, stability and accuracy have been significantly improved. In addition, the novel algorithm is used in the parameter optimization of support vector machines (ULWPSOSA-SVM), and the experimental results indicate that it gets a better classification performance compared with SVM and PSO-SVM.


2013 ◽  
Vol 394 ◽  
pp. 505-508 ◽  
Author(s):  
Guan Yu Zhang ◽  
Xiao Ming Wang ◽  
Rui Guo ◽  
Guo Qiang Wang

This paper presents an improved particle swarm optimization (PSO) algorithm based on genetic algorithm (GA) and Tabu algorithm. The improved PSO algorithm adds the characteristics of genetic, mutation, and tabu search into the standard PSO to help it overcome the weaknesses of falling into the local optimum and avoids the repeat of the optimum path. By contrasting the improved and standard PSO algorithms through testing classic functions, the improved PSO is found to have better global search characteristics.


2012 ◽  
Vol 532-533 ◽  
pp. 1664-1669 ◽  
Author(s):  
Jun Li Zhang ◽  
Da Wei Dai

For the purpose of overcoming the premature property and low execution efficiency of the Particle Swarm Optimization (PSO) algorithm, this paper presents a particle swarm optimization algorithm based on the pattern search. In this algorithm, personal and global optimum particles are chosen in every iteration by a probability. Then, local optimization will be performed by the pattern search and then the original individuals will be replaced. The strong local search function of the pattern search provides an effective mechanism for the PSO algorithm to escape from the local optimum, which avoids prematurity of the algorithm. Simulation shows that this algorithm features a stronger function of global search than conventional PSO, so that the optimization process can be improved remarkably.


2018 ◽  
Vol 18 (3) ◽  
pp. 62-74 ◽  
Author(s):  
Romasevych Yuriy ◽  
Loveikin Viatcheslav

Abstract Since canonical PSO method has many disadvantages which do not allow to effectively reach the global minima of various functions it needs to be improved. The article refers to a novel Multi-Epoch Particle Swarm Optimization (ME-PSO) technique which has been developed by authors. ME-PSO algorithm is based on reinitializing of the stagnant swarm with low exploration efficiency. This approach provides a high rate of global best changing. As a result ME-PSO has great possibility of finding good local (or even global) optimum and does not trap in bad local optimum. In order to prove the advantages of the ME-PSO technique numerical experiments have been carried out with ten uni- and multimodal benchmark functions. Analysis of the obtained results convincingly showed significant superiority of ME-PSO over PSO and IA-PSO algorithms. It has been set that canonical PSO is a special case of ME-PSO.


Sign in / Sign up

Export Citation Format

Share Document