基于市场的再制造P制造系统集成库存随机最优控制研究

在逆向物流中,再制造零部件的数量具有不确定性.根据这一特点,将其看成是随机参数,给出了再制造/制造系统集成库存模式.基于市场对再生品和新产品的不同需求,利用马尔科夫决策理论对具有随机再制造零部件的再制造/制造系统集成库存进行研究,考虑再制造零部件单位成本、新零部件(新购或新制造)单位成本和固定成本及有缺货赔偿的情况,给出了马尔科夫决策过程模型,得到了随机最优控制策略,并用案例进行分析验证.     

MDP平均模型的强最优性

考虑的是任意状态空间，任意行动空间ＭＤＰ平均模型的四个平均准则，在Ｏ．Ｈ．Ｌｅｒｍａ的遍历性条件下，利用稳定性定理和可测选择理论简明地证明了存在平稳策略关于此模型的四个平均准则同时是最优的，从而扩充并加强了Ｏ．Ｈ．Ｌｅｒｍａ（１９８９）的主要结果。     

基于强化学习的写字楼动态电力价格策略研究

针对城市商业区写字楼的微电网能源管理系统,如何有效的进行能量协同调度具有重要的理论价值与应用价值。本文提出了基于强化学习的写字楼能量调度的需求侧动态价格响应算法。首先,将物业与写字楼之间的模型建立为一个马尔可夫决策过程（Markov Decision Process,MDP）,通过使用Q-Learning算法中的贪婪策略（ε - greedy）来探索写字楼中每个时间间隙的最佳电力零售价格。实验表明,通过需求侧动态价格能量调度策略不仅可以提高电网的可靠性和能源的有效利用率,实现电力分流、削峰和填谷,也可有效降低写字楼的能量消耗,同时降低用户的电力能耗使用成本,使得各方利益最大化,验证了本文所提算法的有效性和实用性。     

异构网络中基于最小二乘权值法的MDP垂直切换算法研究

文中将多属性QoS考虑在内,针对业务特点,提出一种基于最小二乘(LS,Least Square)权值法的MDP(Markov decision processes,马尔可夫决策过程)(LS_MDP)垂直切换算法.为了评估性能,考虑3GPP定义的两类业务,根据业务特点,利用LS法获得决策者对指标的权重向量;接着利用MDP分析方法进行垂直切换判决.通过对算法的性能分析,可以看出LS_MDP算法能够根据各业务特点综合考虑各属性间关系,获得优良的系统性能.     

Leishmania donovani in hamsters: Stimulation of non-specific resistance by some novel glycopeptides and impact on therapeutic efficacy

Summary Several glycopeptides structurally related to muramyl dipeptide (MDP) have been synthesized and evaluated for their ability to stimulate the non-specific resistance of hamsters againstL. donovani infection. These compounds have been named CDRI compounds. The synthetic procedure used for compounds 86/448 and 84/212 is described.MDP and its synthetic congeners were administered as immunostimulants at a prophylactic dose of 3 mg/kg at two weeks interval. The challenge infection (1×10⁷ amastigotes i.c./hamster) was given in between two doses of the compounds. One of the glycopeptides, CDRI comp. 86/448, has been found to be significantly more potent than MDP, effecting 92% inhibition of the challenge dose, whereas MDP produced only 26.5% inhibition. The effect of comp. 86/448 lasted until day 7 of challenge. The efficacy of sodium stibogluconate was appreciably improved in hamsters treated with comp. 86/448.     

Q（f）－过程非唯一时连续时间折扣目标MDP

考虑的是可数状态空间连续时间ＭＤＰ的折扣模型，与以往不同的是，我们放弃了由策略ｆ所确定的Ｑ（ｆ）－过程唯一的传统假设．而首次考虑Ｑ（ｆ）－过程非唯一的情形，借助于Ｑ－过程的构造理论，用拓扑分析的方法，证明了最优策略的存在性．     

求解有限马氏决策问题的神经网络方法

在有限马尔可夫决策过程的线性规划求解方法以及神经网络算法的基础上提出了运用神经网络求解有限马尔可夫决策问题的方法．并通过算例验证了该方法的有效性．     

Leishmania donovani in hamsters: Stimulation of non-specific resistance by novel lipopeptides and their effect in antileishmanial therapy

Several novel type of lipopeptides were synthesized and evaluated for their ability to stimulate non-specific resistance againstLeishmania donovani infection. Peritoneal macrophages isolated from young male hamsters treated with muramyl dipeptide (MDP) and various synthetic lipopeptides (6 mg/kg i.p.) 7 days earlier, were cultured in vitro and challenged 24 h later withL. donovani promastigotes. One lipopeptide, Central Drug Research Institute (CDRI) compound 86/450, exhibited significantly higher immunostimulatory activity than MDP. Its prophylactic activity was further confirmed in hamsters by giving 2 split doses of 3 mg/kg of the compound spaced at 2 weeks, i.e. on day –7 and +7 of challenge withL. donovani amastigotes. The prophylactic effect lasted for 7 days following the last treatment with compound 86/450. The antileishmanial action of sodium stibogluconate (SAG) was also found to be enhanced by 16% in hamsters primed with compound 86/450.CDRI Communication No. 5034.     

Fuzzy Q learning algorithm for dual-aircraft path planning to cooperatively detect targets by passive radars

The problem of passive detection discussed in this paper involves searching and locating an aerial emitter by dualaircraft using passive radars. In order to improve the detection probability and accuracy, a fuzzy Q learning algorithrn for dual-aircraft flight path planning is proposed. The passive detection task model of the dual-aircraft is set up based on the partition of the target active radar＇s radiation area. The problem is formulated as a Markov decision process （MDP） by using the fuzzy theory to make a generalization of the state space and defining the transition functions, action space and reward function properly. Details of the path planning algorithm are presented. Simulation results indicate that the algorithm can provide adaptive strategies for dual-aircraft to control their flight paths to detect a non-maneuvering or maneu- vering target.