基于博弈分析的自动协商系统

自动协商作为一个热点已经研究了很多年。大多数研究工作都着重于研究独立协商应用的抽象和理论模型,而对于实际算法的应用性只做了很少的工作。主要提出了一种基于博弈论的比较有效的协商模型来解决协商中的冲突。在该模型中利用遗传算法进行策略优化,而利用另外一个算法对已有的No-Fear-of-Deviation（NFD）算法进行了改进。     

A negotiation methodology for multidisciplinary collaborative product design

The collaborative design of a complicated mechanical product often involves conflicting multidisciplinary objectives, thus one key problem is conflict resolution and coordination among the different disciplines. Since the characteristics such as cooperative competition, professional dependence, compromise, overall utility and so on exist in multidisciplinary collaborative design (MCD), an effective way to gradually eliminate the conflicts among the multiple disciplines and reach an agreement is the negotiation by which a compromise solution that satisfies all parties is got. By comprehensively analyzing the characteristics in MCD and considering the benefit equilibrium among discipline individuals and team, a negotiation strategy is presented, which maximize the union satisfaction degree of system overall objective under the premise of ensuring the higher satisfaction degree level of each discipline’s local objective. A design action of a discipline is abstractly expressed as a concession in the negotiation strategy, and a negotiation model used for MCD is generated by establishing the relation between concession and satisfaction degree. By the relation between satisfaction degree and objective function, the mapping relationship between satisfaction degree domain and physical domain is built to get the design solution. A negotiation process is planned, and a negotiation system framework is designed to support the negotiation among multiple disciplines and assist the different disciplines rapidly reach a consistent compromise solution. A design example of automotive friction clutch is given to illustrate the proposed method.     

基于博弈论及Q学习的多Agent协作追捕算法

多Agent协作追捕问题是多Agent协调与协作研究中的一个典型问题。针对具有学习能力的单逃跑者追捕问题，提出了一种基于博弈论及Q学习的多Agent协作追捕算法。首先,建立协作追捕团队，并构建协作追捕的博弈模型；其次,通过对逃跑者策略选择的学习，建立逃跑者有限的Step-T累积奖赏的运动轨迹，并把运动轨迹调整到追捕者的策略集中；最后,求解协作追捕博弈得到Nash均衡解，每个Agent执行均衡策略完成追捕任务。同时,针对在求解中可能存在多个均衡解的问题，加入了虚拟行动行为选择算法来选择最优的均衡策略。C#仿真实验表明，所提算法能够有效地解决障碍环境中单个具有学习能力的逃跑者的追捕问题，实验数据对比分析表明该算法在同等条件下的追捕效率要优于纯博弈或纯学习的追捕算法。     

一种军棋机器博弈的多棋子协同博弈方法

针对在军棋博弈不完全信息对弈中,面对棋子不同价值、不同位置、不同搭配所产生的不同棋力,传统的单子意图搜索算法,不能满足棋子之间的协同性与沟通性,同时也缺乏对敌方的引诱与欺骗等高级对抗能力。本文提出一种结合UCT搜索策略的高价值棋子博弈方法,实现高价值棋子协同博弈的策略。实战经验表明:高价值多棋子军棋协同博弈策略优于单棋子军棋博弈策略。     

A multi-agent platform for content-based image retrieval

Efficient and possibly intelligent image retrieval is an important task, often required in many fields of human activity. While traditional database indexing techniques exhibit a remarkable performance in textual information retrieval current research in content-based image retrieval is focused on developing novel techniques that are biologically motivated and efficient. It is well known that humans have a remarkable ability to process visual information and to handle the volume and complexity of such information quite efficiently. In this paper, we present a content-based image retrieval platform that is based on a multi-agent architecture. Each agent is responsible for assessing the similarity of the query image to each candidate image contained in a collection based on a specific primitive feature and a corresponding similarity criterion. The outputs of various agents are integrated using one of several voting schemes supported by the system. The system’s performance has been evaluated using various collections of images, as well as images obtained in specific application domains such as medical imaging. The initial evaluation has yielded very promising results.

A protocol for multi-agent negotiation in a group-choice decision making process

Negotiation is one of the most important features of agent interactions found in multi-agent systems, because it provides the basis for managing the expectations of the individual negotiating agents, and it enables selecting solutions that satisfy all the agents as much as possible. In order for negotiation to take place between two or more agents there is need for a negotiation protocol that defines the rules of the game; consequently, a variety of agent negotiation protocols have been proposed in literature. However, most of them are inappropriate for Group-Choice Decision Making (GCDM) because they do not explicitly exploit tradeoff to achieve social optimality, and their main focus is solving two-agent negotiation problems such as buyer–seller negotiation. In this paper we present an agent negotiation protocol that facilitates the solving of GCDM problems. The protocol is based on a hybrid of analytic and artificial intelligence techniques. The analytic component of the protocol utilizes a Game Theory model of an n-person general-sum game with complete information to determine the agreement options, while the knowledge-based (artificial intelligence) component of the protocol is similar to the strategic negotiation protocol. Moreover, this paper presents a tradeoff algorithm based on Qualitative Reasoning, which the agents employ to determine the ‘amount’ of tradeoff associated with various agreement options. Finally, the paper presents simulation results that illustrate the operational effectiveness of our agent negotiation protocol.     

面向多智能体博弈的并行蒙特卡洛树搜索算法研究

蒙特卡洛树搜索算法是一种常用的强化学习算法,博弈过程中动态空间的指数级增长是制约该算法学习效率的因素。基于并行方法对蒙特卡洛树搜索算法进行优化,提出基于胜率估值传递的并行蒙特卡洛树搜索算法。改进后的并行博弈搜索策略框架包含一个主进程和多个子进程,其中子进程用于探索,主进程根据子进程传递的胜率估值数据进行决策。结合多智能体博弈平台Pommerman进行实验验证,与传统的蒙特卡罗树搜索算法相比,并行蒙特卡罗树搜索算法有效提高了资源利用率、博弈胜率及决策效率。     

A multi-agent system performing one-to-many negotiation for load balancing of electricity use

Emerging technologies allowing two-way communication between utility companies and their customers are changing the rules of the energy market. Deregulation makes it even more demanding for utility companies to create new business processes for the mutual benefit of the companies and their customers. Dynamic load management of the power grid is essential to make better and more cost-effective use of electricity production capabilities, and to increase customer satisfaction. In this paper, methods from agent technology and knowledge technology have been used to analyse, design, and implement a component-based multi-agent system capable of negotiation for load management. The proof-of-concept prototype system NALM (negotiating agents for load management) developed shows how under certain assumptions peaks in power load can be reduced effectively based on a negotiation process.     

A multi-agent intelligent system for efficient ERP maintenance

The Enterprise Resource Planning (ERP) system is an enterprise-wide integrated software package designed to uphold the highest quality standards of business process. However, for the time being, when the business condition has been changed, the system may not guarantee that the process embedded in ERP is still best. Moreover, since the ERP system is very complex, maintaining the system by trial and error is very costly. Hence, this paper aims to construct a support system that adjusts ERP system to environmental changes. To do so, we adopt multi-agent intelligent technology that enables autonomous cooperation with one another to monitor ERP databases and to find any exceptional changes and then analyze how the changes will affect ERP performance. Moreover, Petri net is applied to manage the complexity and dynamics of agents’ behavior. To show the feasibility of the idea, a prototype agent system, ERP/PN, is proposed and an experiment is conducted.     

A multi-agent system for distributed multi-project scheduling: An auction-based negotiation approach

Simultaneously running multiple projects are quite common in industries. These projects require local (always available to the concerned project) and global (shared among the projects) resources that are available in limited quantity. The limited availability of the global resources coupled with compelling schedule requirements at different projects leads to resource conflicts among projects. Effectively resolving these resource conflicts is a challenging task for practicing managers. This paper proposes a novel distributed multi-agent system using auctions based negotiation (DMAS/ABN) approach for resolving the resource conflicts and allocating multiple different types of shared resources amongst multiple competing projects. The existing multi-agent system (MAS) using auction makes use of exact methods (e.g. dynamic programming relaxation) for solving winner determination problem to resolve resource conflicts and allocation of single unit of only one type of shared resource. Consequently these methods fail to converge for some multi-project instances and unsuitable for real life large problems. In this paper the multi-unit combinatorial auction is proposed and winner determination problem is solved by efficient new heuristic.The proposed approach can solve complex large-sized multi-project instances without any limiting assumptions regarding the number of activities, shared resources or the number of projects. Additionally our approach further allows to random project release-time of projects which arrives dynamically over the planning horizon. The DMAS/ABN is tested on standard set of 140 problem instances. The results obtained are benchmarked against the three state-of-the-art decentralized algorithms and two existing centralized methods. For 82 of 140 instances DMAS/ABN found new best solutions with respect to average project delay (APD) and produced schedules on an average 16.79% (with maximum 57.09%) lower APD than all the five methods for solving the same class of problems.     

A fuzzy-based negotiation approach for collaborative planning in manufacturing supply chains

The coordination of the planning operations across the manufacturing supply chains (MSC) is considered as a major component of supply chain management. As centralized coordination requires relevant information sharing, alternative approaches are needed to synchronize production plans between partners of MSC characterized by decentralized decision making systems with limited information sharing. In this paper, a bi-level fuzzy-based negotiation approach is proposed in order to model collaborative planning between MSC partners. During negotiation, each manufacturer is optimizing a bi-objective planning model. In order to generate optimal production plans, a genetic algorithm is used. To evaluate the exchanged proposals and the satisfaction degree of each partner, the fuzzy logic approach is adopted in the both negotiation levels. The main result of the developed approach consists in a collaborative decision making mechanism allowing the MSC partners to define their optimal production plans while considering the whole negotiating process with the pre-negotiation and post-negotiation stages. Computational tests done for different MSC structures show the effectiveness of the proposed mechanism, which ensures the satisfaction of the manufacturers and the optimality of the final solution. By comparing the results with the ones obtained considering a centralized planning approach, it is shown that the developed negotiation mechanism yields to near optimal solutions with insignificant gaps from the global optimal solutions.     

基于多代理的供应链协同生产计划模型

供应链是由分布在全球的供应商、制造商、仓库、分销中心和零售商组成的复杂网络。其生产计划具有分布性,自治性,同步性和开放性等特点,使得传统的生产计划方法已经不能适应供应链的计划需求。在分析了多代理技术和供应链系统生产计划特点的基础上,采用智能代理封供应链系统的功能实体和物理实体进行封装,提出了一个基于多代理的供应链系统网络模型,并构建了基于多智能代理的生产计划运行模式。该生产计划模型分为三层：全局生产计划,企业内部子生产计划和各个任务的详细生产计划。它突破了传统生产计划的局限性,从全局规划的角度来整合供应链上的所有资源,消除了不同企业子生产计划所产生的冲突和差异。很好的体现了供应链系统信息共享和资源共享的原则。     

A virtual collaborative maintenance architecture for manufacturing enterprises

This paper presents a virtual collaborative maintenance architecture aimed at improving the performance of manufacturing systems. The proposed architecture incorporates maintenance elements such as operational reliability, maintenance economics, human factors in maintenance, maintenance program, and maintenance optimization in a virtual collaborative architecture. An analytical model is proposed to measure the relative performance of the proposed virtual collaborative architecture as well as that of the manufacturing enterprise. A numerical example is also presented to demonstrate the application of the proposed approach.     

A game theoretical model for collaborative groups in social applications

In this paper, we address the problem of the free riding behaviour that takes advantage of collaborative educational social groups without contributing back to other participants posts. Free riders are active users who ask questions and draw knowledge from the community but provide very limited or no contributions back to it. Since the survival of a collaborative educational community is highly dependent on its active users and their contributions, motivating free riding users to take an active part would naturally augment the value the community provides and ensure its survivability. As a solution, we formally analyse the impact of the free riding behaviour by means of repeated game theory where classical and generous Tit for Tat are used. Such analysis shows the impact of such behaviour on educational communities and raises the need for other strategies that motivate free riding users to cooperate under the threat of being punished by cooperative ones; hence, we introduce reputation based Tit for Tat strategies. Our study suggests adding reputation as a parameter in users’ profiles in collaborative groups to improve their survivability.     

煤矿井下基于博弈论CR节点协作通信研究

现有无线通信系统在不同矿区通用性不强、通信效果不佳。煤矿井下认知系统可通过对煤矿井下通信环境参数的感知,对感知参数信息的融合、分析,利用动态博弈理论实现动态的频率选择、功率分配、调制方式等参数的调整,获得有效的通信效果和不同矿区兼容。通过Matlab进行了仿真验证,仿真结果表明：该算法能够快速收敛于系统稳定状态,而且,复杂度低,易于实现。     

A collaborative design platform for new alloy material development

To overcome the shortcomings of the conventional trial and error mode for new material development, a full-process collaborative design platform for steel rolling is developed based on an industrial internet of things (IIoT) system in this study. Equipment, process and product entities are modeled in both the physical domain and the cyber domain. A systematic data-driven Mamdani-type fuzzy modeling methodology is proposed to map the relationship between material chemical compositions, organizational structures, process parameters and mechanical performances. The proposed methodology employs a random forest (RF) algorithm to select important parameters from mechanism models, simulation models and production process variables, utilizes a K-means algorithm to merge diverse steel grades into sub-clusters, and implements a multi-objective particle swarm optimization (MOPSO) algorithm to further improve the fuzzy model in terms of both the structure and the membership function parameters. A dataset of 3500 steel coils collected by the prototype platform built in a large hot rolling mill is used to evaluate the performance of the proposed approach. Experiment results show that the proposed methodology performs well in predicting the yield strength, tensile strength and elongation, with the coverage probability over 90% under 10% deviation and about 70% under 5% deviation on average.     

A platform for P2P agent-based collaborative applications

The operational environment can be a valuable source of information about the behavior of software applications and their usage context. Although a single instance of an application has limited evidence of the range of the possible behaviors and situations that might be experienced in the field, the collective knowledge composed by the evidence gathered by the many instances of a same application running in several diverse user environments (eg, a browser) might be an invaluable source of information. This information can be exploited by applications able to autonomously analyze how they behave in the field and adjust their behavior accordingly. Augmenting applications with the capability to collaborate and directly share information about their behavior is challenging because it requires the definition of a fully decentralized and dependable networked infrastructure whose nodes are the user machines. The nodes of the infrastructure must be collaborative, to share information, and autonomous, to exploit the available information to change their behavior, for instance, to better accommodate the needs of the users to prevent known problems. This paper describes the initial results that we obtained with the design and the development of an infrastructure that can enable the execution of collaborative scenarios in a fully decentralized way. Our idea is to combine the agent-based paradigm, which is well suited to design collaborative and autonomous nodes, and the peer-to-peer paradigm, which is well suited to design distributed and dynamic network infrastructures. To demonstrate our idea, we augmented the popular JADE agent-based platform with a software layer that supports both the creation of a fully decentralized peer-to-peer network of JADE platforms and the execution of services within that network, thus enabling JADE multiagent systems (MASs) to behave as peer-to-peer networks. The resulting platform can be used to study the design of collaborative applications running in the field.     

A service-oriented city portal framework and collaborative development platform

Cities are being equipped with multiple information systems to provide public services for city officials, officers, citizens, and tourists. There have been concerns with efficient service implementation and provision, e.g., data islands and function overlaps between systems and applications. Service-oriented portals are efficient at facilitating information sharing and collaborative work between city systems and users. The goal of this research is to make cities responsive, agile and to provide composite services efficiently and cost efficiently. A service-oriented framework for city portals is proposed to design, integrate and streamline city systems and applications. A model driven collaborative development platform of the proposed framework was developed for service-oriented digital portals. The architecture and implementation issues of the platform are discussed. The service identification policies are discussed within the framework. A case study has been developed and evaluated on the platform to provide a composite service, i.e., a traffic search service on a city portal.     

A multi-agent platform for auction-based allocation of loads in transportation logistics

This paper describes an agent-based platform for the allocation of loads in distributed transportation logistics, developed as a collaboration between CWI, Dutch National Center for Mathematics and Computer Science, Amsterdam and Vos Logistics Organizing, Nijmegen, The Netherlands.The platform follows a real business scenario proposed by Vos, and it involves a set of agents bidding for transportation loads to be distributed from a central depot in the Netherlands to different locations across Germany. The platform supports both human agents (i.e. transportation planners), who can bid through specialized planning and bidding interfaces, as well as automated, software agents. We exemplify how the proposed platform can be used to test both the bidding behaviour of human logistics planners, as well as the performance of automated auction bidding strategies, developed for such settings.The paper first introduces the business problem setting and then describes the architecture and main characteristics of our auction platform. We conclude with a preliminary discussion of our experience from a human bidding experiment, involving Vos planners competing for orders both against each other and against some (simple) automated strategies.     

Multi-agent team cooperation: A game theory approach

The main goal of this work is to design a team of agents that can accomplish consensus over a common value for the agents’ output using cooperative game theory approach. A semi-decentralized optimal control strategy that was recently introduced by the authors is utilized that is based on minimization of individual cost using local information. Cooperative game theory is then used to ensure team cooperation by considering a combination of individual cost as a team cost function. Minimization of this cost function results in a set of Pareto-efficient solutions. Among the Pareto-efficient solutions the Nash-bargaining solution is chosen. The Nash-bargaining solution is obtained by maximizing the product of the difference between the costs achieved through the optimal control strategy and the one obtained through the Pareto-efficient solution. The latter solution results in a lower cost for each agent at the expense of requiring full information set. To avoid this drawback some constraints are added to the structure of the controller that is suggested for the entire team using the linear matrix inequality (LMI) formulation of the minimization problem. Consequently, although the controller is designed to minimize a unique team cost function, it only uses the available information set for each agent. A comparison between the average cost that is obtained by using the above two methods is conducted to illustrate the performance capabilities of our proposed solutions.