An improved particle swarm optimization (PSO) algorithm based on ensemble technique is presented. The algorithm combines some previous best positions (pbest) of the particles to get an ensemble position (Epbest), whic...An improved particle swarm optimization (PSO) algorithm based on ensemble technique is presented. The algorithm combines some previous best positions (pbest) of the particles to get an ensemble position (Epbest), which is used to replace the global best position (gbest). It is compared with the standard PSO algorithm invented by Kennedy and Eberhart and some improved PSO algorithms based on three different benchmark functions. The simulation results show that the improved PSO based on ensemble technique can get better solutions than the standard PSO and some other improved algorithms under all test cases.展开更多
The previous particle swarm optimizers lack direct mechanism to prevent particles beyond predefined search space, which results in invalid solutions in some special cases. A momentum factor is introduced into the orig...The previous particle swarm optimizers lack direct mechanism to prevent particles beyond predefined search space, which results in invalid solutions in some special cases. A momentum factor is introduced into the original particle swarm optimizer to resolve this problem. Furthermore, in order to accelerate convergence, a new strategy about updating velocities is given. The resulting approach is mromentum-PSO which guarantees that particles are never beyond predefined search space without checking boundary in every iteration. In addition, linearly decreasing wight PSO （LDW-PSO） equipped with a boundary checking strategy is also discussed, which is denoted as LDWBC-PSO. LDW-PSO, LDWBC-PSO and momentum-PSO are compared in optimization on five test functions. The experimental results show that in some special cases LDW-PSO finds invalid solutions and LDWBC-PSO has poor performance, while momentum-PSO not only exhibits good performance but also reduces computational cost for updating velocities.展开更多
A dynamic parallel forecasting model is proposed, which is based on the problem of current forecasting models and their combined model. According to the process of the model, the fuzzy C-means clustering algorithm is ...A dynamic parallel forecasting model is proposed, which is based on the problem of current forecasting models and their combined model. According to the process of the model, the fuzzy C-means clustering algorithm is improved in outliers operation and distance in the clusters and among the clusters. Firstly, the input data sets are optimized and their coherence is ensured, the region scale algorithm is modified and non-isometric multi scale region fuzzy time series model is built. At the same time, the particle swarm optimization algorithm about the particle speed, location and inertia weight value is improved, this method is used to optimize the parameters of support vector machine, construct the combined forecast model, build the dynamic parallel forecast model, and calculate the dynamic weight values and regard the product of the weight value and forecast value to be the final forecast values. At last, the example shows the improved forecast model is effective and accurate.展开更多
现实工业生产应用中存在大量的混合变量优化问题,这类问题的决策变量既包含连续变量,又包含离散变量。由于决策变量为混合类型,导致问题的决策空间变得不规则,采用已有的方法很难进行有效求解。引入协同进化策略,提出一种基于竞争式协...现实工业生产应用中存在大量的混合变量优化问题,这类问题的决策变量既包含连续变量,又包含离散变量。由于决策变量为混合类型,导致问题的决策空间变得不规则,采用已有的方法很难进行有效求解。引入协同进化策略,提出一种基于竞争式协同进化的混合变量粒子群优化算法(competitive coevolution based PSO,CCPSO)。设计基于容忍度的搜索方向调整机制来判断粒子的进化状态,从而自适应地调整粒子的搜索方向,避免陷入局部最优,平衡了种群的收敛性和多样性;引入基于竞争式协同进化的学习对象生成机制,在检测到粒子进化停滞时为每个粒子生成新的学习对象,从而推动粒子的进一步搜索,提高了种群的多样性;采用基于竞争学习的预测策略为粒子选择合适的学习对象,充分利用了新旧学习对象的学习潜力,保证了算法的收敛速度。实验结果表明:相比其他主流的混合变量优化算法,CCPSO可以获得更优的结果。展开更多
文摘An improved particle swarm optimization (PSO) algorithm based on ensemble technique is presented. The algorithm combines some previous best positions (pbest) of the particles to get an ensemble position (Epbest), which is used to replace the global best position (gbest). It is compared with the standard PSO algorithm invented by Kennedy and Eberhart and some improved PSO algorithms based on three different benchmark functions. The simulation results show that the improved PSO based on ensemble technique can get better solutions than the standard PSO and some other improved algorithms under all test cases.
文摘The previous particle swarm optimizers lack direct mechanism to prevent particles beyond predefined search space, which results in invalid solutions in some special cases. A momentum factor is introduced into the original particle swarm optimizer to resolve this problem. Furthermore, in order to accelerate convergence, a new strategy about updating velocities is given. The resulting approach is mromentum-PSO which guarantees that particles are never beyond predefined search space without checking boundary in every iteration. In addition, linearly decreasing wight PSO （LDW-PSO） equipped with a boundary checking strategy is also discussed, which is denoted as LDWBC-PSO. LDW-PSO, LDWBC-PSO and momentum-PSO are compared in optimization on five test functions. The experimental results show that in some special cases LDW-PSO finds invalid solutions and LDWBC-PSO has poor performance, while momentum-PSO not only exhibits good performance but also reduces computational cost for updating velocities.
基金supported by the National Defense Preliminary Research Program of China(A157167)the National Defense Fundamental of China(9140A19030314JB35275)
文摘A dynamic parallel forecasting model is proposed, which is based on the problem of current forecasting models and their combined model. According to the process of the model, the fuzzy C-means clustering algorithm is improved in outliers operation and distance in the clusters and among the clusters. Firstly, the input data sets are optimized and their coherence is ensured, the region scale algorithm is modified and non-isometric multi scale region fuzzy time series model is built. At the same time, the particle swarm optimization algorithm about the particle speed, location and inertia weight value is improved, this method is used to optimize the parameters of support vector machine, construct the combined forecast model, build the dynamic parallel forecast model, and calculate the dynamic weight values and regard the product of the weight value and forecast value to be the final forecast values. At last, the example shows the improved forecast model is effective and accurate.
文摘现实工业生产应用中存在大量的混合变量优化问题,这类问题的决策变量既包含连续变量,又包含离散变量。由于决策变量为混合类型,导致问题的决策空间变得不规则,采用已有的方法很难进行有效求解。引入协同进化策略,提出一种基于竞争式协同进化的混合变量粒子群优化算法(competitive coevolution based PSO,CCPSO)。设计基于容忍度的搜索方向调整机制来判断粒子的进化状态,从而自适应地调整粒子的搜索方向,避免陷入局部最优,平衡了种群的收敛性和多样性;引入基于竞争式协同进化的学习对象生成机制,在检测到粒子进化停滞时为每个粒子生成新的学习对象,从而推动粒子的进一步搜索,提高了种群的多样性;采用基于竞争学习的预测策略为粒子选择合适的学习对象,充分利用了新旧学习对象的学习潜力,保证了算法的收敛速度。实验结果表明:相比其他主流的混合变量优化算法,CCPSO可以获得更优的结果。
文摘粒子群优化(particle swarm optimization,PSO)算法是一种在机器人运动规划、信号处理等领域有广泛应用的优化算法。然而该算法易陷入局部最优解,从而导致早熟问题。出现早熟问题的原因之一是粒子群仅依靠适应度值选择学习范例。为了克服上述问题,提出了一种基于适应度值、改进率和新颖性混合驱动的PSO算法(particle swarm optimization algorithm based on hybrid driven by fitness values,improvement rate,and novelty,FINPSO)。在该算法中,引入的新指标和遗传算法会平衡种群的探索与开发,降低粒子群早熟的可能性。适应度值、改进率和新颖性会作为粒子的评价指标。各指标独立地选择学习范例并保存到不同的档案中。粒子每一次速度更新都要确定各个指标的权重,并从每个档案中选择一个范例学习。该算法采用了遗传算法进行粒子间的信息交流。遗传算法中的交叉互换和突变会给种群带来更多的随机性,提升种群的全局搜索能力。以八个PSO算法变体作为对比算法,两个CEC测试套件作为基准函数进行实验。实验结果表明,FINPSO算法优于已有的PSO算法变体达到最先进水平。
