基于多Agent的季节性商品动态定价算法

研究两个提供商销售季节性商品时的最优定价策略问题。在性能势理论的基础上,针对季节性商品的特殊属性,建立两个提供商之间没有信息交互情况下的季节性商品的动态定价模型,并引入了Q学习算法和Wolf-PHC算法。通过仿真实验对DF方法定价,Q学习算法定价和Wolf-PHC算法定价进行比较,得到Wolf-PHC算法定价的优化效果更明显,适应性更强。     

考虑时变奖励的多摇臂算法在动态定价中的应用

考虑到动态定价是一个非固定性的多摇臂（Multi-Armed Bandit,MAB）问题,即厂商的利润会随时间变化,因此在相关研究基础上,研究了需求不确定情况下考虑时变奖励的置信区间上界（Upper Confidence Bound,UCB）算法在动态定价问题上的应用。将商品定价问题描述为一个多摇臂问题,并构建利润最大化模型求得最优解。仿真结果表明,通过将考虑时变奖励的置信区间上界算法与基础的多摇臂算法进行对比分析,所提出的算法学得的奖励更加接近真实奖励,收敛速度更快。相较于前人研究,该模型考虑了时变因素,更加符合现实场景中的动态定价,为厂商定价提供了相应的决策支持。     

粒子群算法在季节性商品最优定价中的应用

从最小化期望损失的角度建立了季节性商品的最优定价模型，并采用粒子群算法进行求解。结合具体算例，根据不同库存量、库存量和折扣价的不同组合，分别获得达到最小期望损失的最优定价，可以很好地解释模型所具有的经济意义。对仿真结果的分析表明：粒子群算法不仅能灵活、简便地获得多种情况下的最优定价，而且反映了最优定价在库存量和折扣价不同组合时的变化规律，从而为销售商确定最优价格提供建议。因此，应用粒子群算法求解季节性商品的最优定价具有一定的现实意义。     

易腐生鲜货品车辆路径问题的改进混合蝙蝠算法

针对配送易腐生鲜货品的车辆其配送路径的选择不仅受货品类型、制冷环境变化、车辆容量限制、交货时间等多种因素的影响,而且需要达到一定的目标（如：费用最少、客户满意度最高）,构建了易腐生鲜货品车辆路径问题（VRP）的多目标模型,并提出了求解该模型的改进混合蝙蝠算法。首先,采用时间窗模糊化处理方法定义客户满意度函数,细分易腐生鲜货品类型并定义制冷成本,建立了最优路径选择的多目标模型;然后,在分析蝙蝠算法求解离散问题易陷入局部最优、过早收敛等问题的基础上,精简经典蝙蝠算法的速度更新公式,并对混合蝙蝠算法的单多点变异设定选择机制,提高算法性能;最后,对改进混合蝙蝠算法进行性能测试。实验结果表明,与基本蝙蝠算法和已有混合蝙蝠算法相比,所提算法在求解VRP时能够提高客户满意度1.6%~4.2%,且减小平均总成本0.68%~2.91%。该算法具有计算效率高、计算性能好和较高的稳定性等优势。     

量子粒子群算法在易逝品多目标定价中的应用

对易逝品的多目标定价问题进行了研究。从利润最大化角度建立易逝品多目标最优定价模型。模型中涉及复杂的需求函数,常规函数极值法不易获得问题解析解,因此引入量子粒子群算法,结合惩罚函数对模型进行演化求解。根据给出的算例分析表明,利用量子粒子群算法,可以快速有效地得到不同订货量下的最优定价与折扣价组合。     

城市公共交通时间差别定价模型

为解决城市公共交通出行时间需求不均衡的问题,构建了时间差别费率定量计算与复合城市交通网络流量分配双层规划模型。上层为社会福利最优模型,下层通过非必要性出行系数描述多方式分时段复合城市交通网络随机弹性需求的公交时间差别定价模型。运用遗传算法对模型整体进行求解,下层模型采用综合对角化算法和MSA算法的组合求解算法。最后,设计了一个算例以说明模型应用。结果表明,运用时间差别定价方案比高峰/平峰同一费率定价方案社会福利目标函数高37．1,乘客对平峰时段降低票价的弹性高于对高峰时段增加票价的弹性,实行高峰增加票价同时平峰降低票价的方案效果最为明显。     

分布式遗传算法求解TSP问题的CORBA实现研究

论文提出了基于CORBA的分布式遗传算法,描述了实现该算法的三层软件结构模型,并利用该模型对CTSP问题进行了广泛的计算。计算结果表明该模型存在以下三条规律:最优迁移策略与随机迁移策略对最优解影响几乎一致;当迁移率在20‰附近时,易获得最优解;繁衍代数在25000～30000范围内时,易获得最优解。文中还介绍了应用测试数据验证该分布式遗传算法求解最优性的方法。实验表明基于CORBA的分布式遗传算法是求解TSP问题的一种可行、高效的算法。     

基于多摇臂赌博机的产品定价算法

针对在线零售商在不完全需求信息下的单产品定价问题,提出了一种基于多摇臂赌博机的产品定价算法。为了提升多摇臂赌博机算法在定价问题中的效果,该算法利用了需求曲线的单调性,并加入了消费者偏好识别。对消费者的保留价格进行分析得到消费者购买概率,将在线零售商的定价问题建模为多摇臂赌博机模型,给出了相应的定价算法并进行了理论分析,最后通过仿真实验比较了相关算法的定价效果。仿真结果表明该算法提高了在线零售商的收益。     

基于Stackelberg博弈的云CDN缓存资源分配算法

为提高云存储环境下内容分发网络(CDN)的分发效率,提出一种基于Stackelberg博弈的缓存资源分配与定价算法。将Web服务器和云CDN代理商建模成一个多主多从的Stackelberg博弈模型,并构建其各自的效用函数。证明在代理商价格确定的情况下存在Web服务器纳什均衡点,利用一种分布式迭代算法求解博弈模型,得到最优定价与该定价下的最优缓存分配结果。仿真结果表明,该算法可保证Web服务器缓存需求在代理商之间的高效分配,与用户服务质量优先算法相比,可使Web服务器获得更高的单位成本效益。     

多周期多原料的供应链网络成本优化及其混合遗传算法

建立并求解一个基于成本最小的供应链网络模型.与以往研究不同,在该模型中生产一种产品需要至少两种原料,每种原料都可以由备选供应商提供.根据模型的特点,用0、1代表对原材料供应商、工厂和分销中心的选择情况,以MATLAB 7.6为平台,运用Sheffield大学的遗传算法工具箱,将遗传算法与线性规划算法相结合,实现了模型的求解.算例结果表明,给出的染色体编码方案正确,混合遗传算法有效,能解决多周期、多原料的供应链网络成本优化问题.还探讨了需求和距离变化,以及需求随机时对最优成本和最优个体的影响.研究表明,需求变化的影响大于距离变化的影响,需求随机对最优成本和最优个体的影响不大.     

求解TSP问题的模糊自适应粒子群算法

由于惯性权值的设置对粒子群优化（PSO）算法性能起着关键的作用，本文通过引入模糊技术，给出了一种惯性权值的模糊自适应调整模型及其相应的粒子群优化算法，并用于求解旅行商（TSP）问题。实验结果表明了改进算法在求解组合优化问题中的有效性，同时提高了算法的性能，并具有更快的收敛速度。     

基于蚁群信息机制的粒子群算法

针对粒子群算法应用于复杂函数优化时可能出现过早收敛于局部最优解的情况,提出了一种改进的算法。通过构造单个粒子的多个进化方向和类似于蚂蚁群算法信息素表的选择机制,保留了粒子的多种可能进化方向。提高了粒子间的多样性差异,从而改善算法能力。改进后的混合粒子群算法的性能优于带线性递减权重的粒子群算法。     

基于自调节粒子群算法的盲检测

以往基于粒子群优化的盲算法能成功实现信号盲检测,但具有算法收敛速度慢、容易陷入局部最小的缺点。文中通过分析粒子群算法的机能及参数的设置,提出一种改进的基于自调节粒子群优化的盲检测算法。算法构成思想是：基于MIMO系统的盲检测系统模型将盲检测问题转化为二次优化问题,利用改进的自调节粒子群优化算法对此优化问题进行寻优。仿真表明,改进算法具有全局收敛性好、收敛速度快、误码率低的优点,能够更好地解决盲检测问题。     

Particle swarm optimization with query-based learning for multi-objective power contract problem

Particle swarm optimization (PSO) is one of the most important research topics on swarm intelligence. Existing PSO techniques, however, still contain some significant disadvantages. In this paper, we present a new QBL-PSO algorithm that uses QBL (query-based learning) to improve both the exploratory and exploitable capabilities of PSO. Here, we apply a QBL method proposed in our previous research to PSO, and then test this new algorithm on a real case study on problems of power conservation. Our algorithm not only broadens the search diversity of PSO, but also improves its precision. Conventional PSO often snag on local solutions when performing queries, instead of finding better global solutions. To resolve this limitation, when particles converge in nature, we direct some of them into an “ambiguous solution space” defined by our algorithm. This paper introduces two ways to invoke this QBL algorithm. Our experimental results confirm that the proposed method attains better convergence to the global best solution. Finally, we present a new PSO model for solving multi-objective power conservation problems. Overall, this model successfully reduces power consumption, and to our knowledge, this paper represents the first attempt within the literature to apply the QBL concept to PSO.     

A parthenogenetic algorithm for single individual SNP haplotyping

The minimum error correction (MEC) model is one of the important computational models for single individual single nucleotide polymorphism (SNP) haplotyping. Due to the NP-hardness of the model, Qian et al. presented a particle swarm optimization (PSO) algorithm to solve it, and the particle code length is equal to the number of SNP fragments. However, there are hundreds and thousands of SNP fragments in practical applications. The PSO algorithm based on this kind of long particle code cannot obtain high reconstruction rate efficiently. In this paper, a practical heuristic algorithm PGA-MEC based on parthenogenetic algorithm (PGA) is presented to solve the model. A kind of short chromosome code and an effective recombination operator are designed for the algorithm. The reconstruction rate of PGA-MEC algorithm is higher than that of PSO algorithm and the running time of PGA-MEC algorithm is shorter than that of PSO algorithm, which are proved by a number of experiments.     

竞争PSO在多序列联配中的应用

本文提出了一种求解多序列联配的竞争粒子群优化算法,算法根据适应值分类设计了粒子群的惯性权重以及其飞行速度范围,并进行了动态调整,提高了算法的收敛速度和精度;引入了重新初始化机制,有效的避免粒子群优化算法可能出现的早熟现象;提出了一种全新的速度更新模式和竞争策略,增强了算法的能力。实验表明该算法是有效的。     

Common model analysis and improvement of particle swarm optimizer

Particle swarm optimizer （PSO）, a new evolutionary computation algorithm, exhibits good performance for optimization problems, although PSO can not guarantee convergence of a global minimum, even a local minimum. However, there are some adjustable parameters and restrictive conditions which can affect performance of the algorithm. In this paper, the algorithm are analyzed as a time-varying dynamic system, and the sufficient conditions for asymptotic stability of acceleration factors, increment of acceleration factors and inertia weight are deduced. The value of the inertia weight is enhanced to （-1, 1）. Based on the deduced principle of acceleration factors, a new adaptive PSO algorithm- harmonious PSO （HPSO） is proposed. Furthermore it is proved that HPSO is a global search algorithm. In the experiments, HPSO are used to the model identification of a linear motor driving servo system. An Akaike information criteria based fitness function is designed and the algorithms can not only estimate the parameters, but also determine the order of the model simultaneously. The results demonstrate the effectiveness of HPSO.     

Three-objective fuzzy chance-constrained programming model for multiproject and multi-item investment combination

This paper presents a new fuzzy chance-constrained programming model to find the solution for multiproject and multi-item investment combination in investment combination problems. The proposed 0-1 integer programming model has three objectives with fuzzy constraints, and NSGA-II is applied to solve the optimization model with a small modification of the constraint-handling rule. A simulation experiment illustrating the application of the proposed model is presented and Pareto-optimal solutions are obtained through a modified NSGA-II algorithm. A comparison among NSGA-II, PSO, and DE shows that modified NSGA-II has some advantages over PSO and DE.     

Common model analysis and improvement of particle swarm optimizer

Particle swarm optimizer (PSO), a new evolutionary computation algorithm, exhibits good performance for optimization problems, although PSO can not guarantee convergence of a global minimum, even a local minimum. However, there are some adjustable parameters and restrictive conditions which can affect performance of the algorithm. In this paper, the algorithm are analyzed as a time-varying dynamic system, and the sufficient conditions for asymptotic stability of acceleration factors, increment of acceleration factors and inertia weight are deduced. The value of the inertia weight is enhanced to (?1, 1). Based on the deduced principle of acceleration factors, a new adaptive PSO algorithmharmonious PSO (HPSO) is proposed. Furthermore it is proved that HPSO is a global search algorithm. In the experiments, HPSO are used to the model identification of a linear motor driving servo system. An Akaike information criteria based fitness function is designed and the algorithms can not only estimate the parameters, but also determine the order of the model simultaneously. The results demonstrate the effectiveness of HPSO.     

基于文化微粒群算法的丙烯腈收率软测量模型

丙烯腈收率是丙烯腈装置的关键指标,如何得到丙烯腈收率是厂家很关注的研究,将新型优化算法用于丙烯腈收率软测量建模是1种较好的尝试。将新型微粒群优化算法用于同样新型的文化算法种群空间的优化,设计文化微粒群优化算法。它由种群空间和信念空间2部分组成,在种群空间和信念空间分别采用各自算法并行演化,同时,2个空间又根据一定的协议相互联系。分别将该算法和基本微粒群算法用于一些常用测试函数的优化问题;结果表明,与基本微粒群算法相比,文化微粒群算法加强了全局搜索能力,更容易收敛于全局最优解。最后将文化微粒群优化算法用于优化神经网络,构成文化微粒群神经网络,并将其应用于丙烯腈收率软测量建模。结果表明,此模型精度高,应用前景广阔。