MAP segmentation in Bayesian hidden Markov models: a case study

We consider the problem of estimating the maximum posterior probability (MAP) state sequence for a finite state and finite emission alphabet hidden Markov model (HMM) in the Bayesian setup, where both emission and transition matrices have Dirichlet priors. We study a training set consisting of thousands of protein alignment pairs. The training data is used to set the prior hyperparameters for Bayesian MAP segmentation. Since the Viterbi algorithm is not applicable any more, there is no simple procedure to find the MAP path, and several iterative algorithms are considered and compared. The main goal of the paper is to test the Bayesian setup against the frequentist one, where the parameters of HMM are estimated using the training data.     

The competing risks analysis for parallel and series systems using Type-II progressive censoring

Abstract

Recently, the study of the lifetime of systems in reliability and survival analysis in the presence of several causes of failure (competing risks) has attracted attention in the literature. In this paper, series and parallel systems with exponential lifetime for each item of the system are considered. Several causes of failure independently affect lifetime distributions and observations of failure times of the systems are considered under progressive Type-II censored scheme. For series systems, the maximum likelihood estimates of parameters are computed and confidence intervals for parameters of the model are obtained using Fisher information matrix. For parallel systems, the generalized EM algorithm which uses the Newton-Raphson algorithm inside the EM algorithm is used to compute the maximum likelihood estimates of parameters. Also, the standard errors of the maximum likelihood estimates are computed by using the supplemented EM algorithm. The simulation study confirms the good performance of the introduced approach.     

Water cycle algorithm for solving the reliability-redundancy allocation problem with a choice of redundancy strategies

Abstract

This paper proposes a new mathematical model for the reliability-redundancy allocation problem (RRAP) with a choice of redundancy strategies. To maximize the reliability of a system, this model chooses the best redundancy strategy from among both active and standby ones for each subsystem. For those with a standby strategy, a continuous time Markov chain model is used to calculate the exact reliability values. In order to solve the proposed mixed-integer non-linear programing model, a powerful evolutionary algorithm, called water cycle algorithm (WCA), is developed and implemented on three famous benchmark problems. Finally, the results of different benchmark problems are compared with those previously reported to show the superiority of the proposed model and the efficiency of WCA.     

应用混合神经网络和遗传算法的期权价格预测模型

隐含波动率是指在市场中观察的期权价格所蕴涵的波动率.提出了一种加权的隐含波动率作为混合神经网络的输入变量,建立了混合神经网络和遗传算法相结合的期权价格预测模型,通过遗传算法来优化神经网络的结构和获得隐含波动率的权重.在对香港金融衍生品市场的实证中表明,本文模型在预测结果上要优于传统的Black-Scholes模型.     

一类仓库/零售商供应链运输问题的数学模型与算法研究

本文主要研究一类二级供应链中的运输问题,这里是指由若干仓库和零售商组成的供应链,物品从仓库运出,到达客户,以满足客户的需求.分单仓库、多仓库和多产品三种情形分别建立数学模型,设计相应的近似算法,并通过实例讨论了算法的有效性.     

一种求解时变条件下有宵禁限制最短路的算法

在组合优化过程中,往往需要获得从起点到终点之间的最短路.由于道路、天气、交通条件等因素的影响,使得网络具有很强的时变特性.同时,对于网络中的节点往往有宵禁的限制.对时变条件下有宵禁限制并有到达时间限制的最短路进行了研究,建立了软、硬宵禁限制下的数学模型,给出并证明了时变条件下获得有宵禁限制最短路的最优条件,并设计了求解的多项式算法,通过此算法可以获得时变条件下有宵禁限制的最短路.同时,算法和模型还考虑了不同的起点出发时间,使路径决策者可以根据自身的情况,选择合适的出发时间和路径.最后给出了一个应用算例,分析了宵禁对于获得的最短路的影响.     

通货膨胀下的联合库存选址模型研究

研究了通货膨胀对涉及一个供应商、多个候选配送中心和零售商的三级分销网络设计的影响.不同于传统的联合库存选址模型,所研究的模型考虑了整个系统的相关成本的时间价值,其目标是使在一定的通货膨胀率下总的选址、运输和库存成本之和的现值最小.为此,建立了一个非线性的整数规划模型,该模型为无容量约束问题的扩展模型,属于NP-hard问题.提出了3种基于不同编码形式的遗传算法用于求解该模型,并通过不同规模的算例对3种算法进行了有效性的分析和对比.最后,通过对各算例的计算结果的分析,发现随着通货膨胀率的增长,为了使这个分析网络的总成本现值最小,需要增加建立的DC的数量.     

集成化物流管理中库存路径问题研究综述

综合国内外多种参考文献,介绍了IRP的基本概念及其主要特征,从常见模型、求解策略及算法2个维度综述了当前国内外关于IRP的主要研究进展,最后指出了该研究领域在未来研究中应予以重视的几个研究方向,为进一步的研究奠定了基础.     

遗传算法优化神经网络及信用评价研究

研究关于公司神经网络信用评估问题的现状,提出遗传算法辅助网络训练策略(优化后的网络称为进化网络),克服传统网络建模中产生的局部极小缺陷。建立了适合于我国商业企业的信用评分指标体系;然后依据该指标体系建立了基于进化神经网络的信用评估模型;最后,利用样本公司实际指标数据对该模型的评分效果进行了比较研究。     

高等教育最优投资双层规划模型研究

高等教育投资具有"双层"的特点,上层为省级主管部门,下层为高等学校。本文成功运用了双层规划模型,在不考虑高校自筹发展资金投入的情况下,建立了高等教育最优投资双层规划模型,研究了模型最优解的存在性,给出并证明了模型最优解的等价形式,设计了模型解的算法并进行了算法复杂性分析。通过求解模型,可以同时得到省级主管部门和高等学校的最优投资决策方案。文章最后还给出了考虑高校自筹发展资金的两种情况下建立投资模型和求得最优解的方法。