随机博弈

基于随机博弈的Agent协同强化学习方法

A Cooperative Reinforcement Learning Method Based on the Stochastic Game in Multi-Agent Systems

在分析DCRM内涵的基础上，利用随机博弈及ESDP理论，给出了DCRM建模技术及求解思路。

Bases on the analysis of the meaning of dynamic customer relationship management ( DCRM ), it gives the technology of modeling and solving DCRM by combining the stochastic game theory and ESDP .

随机博弈框架下的多agent强化学习方法综述

Survey of Multi-agent Reinforcement Learning in Markov Games

多agent学习是在随机博弈的框架下，研究多个智能体间通过自学习掌握交互技巧的问题。

The research on multi-agent reinforcement learning is to deal with the problem of play skill between agents , just with the concept of stochastic game .

本文针对一类追求系统得益最大化的协作团队的学习问题，基于随机博弈的思想，提出了一种新的多Agent协同强化学习方法。

This paper aims at the learning process of a kind of cooperative teams , which pursue the maximum benefit of a whole system . We propose a new cooperative reinforcement learning method based on the stochastic game in multi-agent systems .

运用随机博弈及可评估的结构动态规划技术建立动态客户关系管理的数学模型。

The dynamic CRM model is presented by using stochastic game theory and estimable structural dynamic programming technologies .

提出了一种刻画网络安全攻防矛盾动态变化，解决最优防御策略选取问题的攻防随机博弈模型。

A new attack-defense stochastic game model is proposed to describe the states dynamically change and address the issue of optimal defense strategy selection .

将网络攻防双方建模为二人非合作攻防随机博弈模型，并详细地描述了该模型的形式化定义及其构成元素。

We model attacker and defender as two players , non-cooperative stochastic game , and detailed describe the formal definition of the model and its component elements .

利用随机博弈及可评估的结构动态规划技术，提出一种使公司及客户利益都最大的最优回报计划的设计方法。

It presents a method of designing the optimal reward programs in a dynamic environment where customers and firm both maximize their benefit by using stochastic game theory and estimable structural dynamic programming technologies .

分析了市场环境下发电商面临的不确定性因素，建立了发电厂出力与报价均衡的随机博弈模型，引入了虚拟机组的概念对能源短缺进行定量计算。

The uncertain factors in electricity market that power generation companies will face with are analyzed , then a stochastic game model is built to realize bidding equilibrium of active power among these companies and a concept of virtual unit is led in to quantitatively calculate the energy shortage .

n人随机重复博弈下的策略选择研究

Research on strategy selection in n-person stochastic iterated game

其三，在Levy扩散环境中，将委托代理问题视为委托人和代理人之间的非零和最优停时-随机控制博弈，其中代理人控制着随机控制过程，委托人选择契约的执行时间。

Finally , when the principal is allowed to choose exercise time in Levy diffusion setting , we formulate the principal-agent problem as a nonzero-sum optimal stopping-stochastic control differential game between the principal and the agent .

基于多目标的随机合作博弈的模糊稳定集

Fuzzy Stable Set of Stochastic Cooperative Games Based on Multi-objective

给出了基于随机攻防博弈模型的最优防御策略选取算法。

An algorithm for defense strategy selection based on those models is proposed .

通过建立一个随机动态博弈模型，考虑债务融资在寡头价格竞争中的策略性效应，并分析市场进入中企业战略性债务融资策略与产品市场竞争策略的选择。

The study result manifests that strategy effect and transaction cost have more influence on adopting vertical integration of the energy firms , while production cost and indetermination make less .

本文对多寡头企业连续时间实物期权投资模型下的能力扩展与随机微分博弈行为及几个关键因素诸如不确定性、最优时机进行分析。

Considering in the same model , the key factors for investment decisions such as uncertainty , flexibility , and optimization timing and the strategies of capacity expansion and Stochastic Differential Games in continuous-time among competitors are analyzed .

首先，在风险股票价格过程服从几何布朗运动的条件下，应用完全平方方法，得到该非零和随机微分博弈问题的显式解，即得到了最优投资组合策略和最优值函数的明确表达式。

Firstly , suppose that the price processes of the risky stock satisfy the geometric Brownian motions . By invoking the use of the completion of square method , we obtain closed-form solutions to the nonzero-sum game problem for the exponential utility function .

并且，基于最小化破产概率准则与基于最大化终止时刻财富期望效用准则的最优投资策略截然不同。（3）基于随机微分博弈方法，研究了存在模型风险时的最优投资决策问题。

Moreover , the optimal strategies are completely different between the case based on maximizing expected utility criteria and based on minimizing ruin probability criteria . ( III ) We study the problem of optimal investment with model risk based on the stochastic differential game approach .

然后，在风险股票价格过程服从马氏调制的几何布朗运动的条件下，应用动态规划原理（HJB方程）得到了该非零和随机微分投资博弈问题的显式解。

Secondly , suppose that the price processes of the risky stock satisfy Markov-modulated geometric Brownian motion . We get the explicit expressions of the optimal portfolio strategies and the optimal value functions for the exponential utility functions by applying the dynamical programming principle ( HJB equations ) .

论第三部门的发展与弱势群体的公众参与强势参与者和弱势参与者群体的随机线性二次博弈

On the Third Sector Development and the Disadvantaged Group 's Participation in Public Affairs Stochastic LQ Games Involving a Major Player and a Large Number of Minor Players

从央视中华民族栏目看强势媒介与弱势群体的共生强势参与者和弱势参与者群体的随机线性二次博弈

Viewing the Intergrowth of Strong Intermediary and weak Groups from the Program the Chinese Nation of CCTV ; Stochastic LQ Games Involving a Major Player and a Large Number of Minor Players

随机结束的重复博弈模型提供了一种新的改善网络团购信用状况的思路：提高双方的重复交易的期望概率，来促使商家选择诚信的策略。

The repeated game model which ended randomly provided a new approach of improving the credit status of group purchasing on internet that is to increase the expected possibility of repeated trades so that sellers will be compelled to choose the credit strategy .

首先，本文将引人中位数来定义随机支苟值的偏好，并在此偏好的基础上进一步定义带随机支付双矩阵博弈的纳什均衡。

In this paper , the preferences on stochastic payoffs are defined by quantiles , and the Nash equilibrium of the bimatrix game with stochastic payoffs is given base on the preferences .