Learning in multi-agent systems: a case study of construction claims negotiation

The ability of agents to learn is of growing importance in multi-agent systems. It is considered essential to improve the quality of peer to peer negotiation in these systems. This paper reviews various aspects of agent learning, and presents the particular learning approach—Bayesian learning—adopted in the MASCOT system (multi-agent system for construction claims negotiation). The core objective of the MASCOT system is to facilitate construction claims negotiation among different project participants. Agent learning is an integral part of the negotiation mechanism. The paper demonstrates that the ability to learn greatly enhances agents' negotiation power, and speeds up the rate of convergence between agents. In this case, learning is essential for the success of peer to peer agent negotiation systems.     

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.     

Evolutionary learning approach to multi-agent negotiation for group recommender systems

Recommender systems (RSs) have emerged as a solution to the information overload problem by filtering and presenting the users with information, services etc. according to their preferences. RSs research has focused on algorithms for recommending items for individual users. However, in certain domains, it may be desirable to be able to recommend items for a group of persons, e.g., movies, restaurants, etc. for which some remarkable group recommender systems (GRSs) have been developed. GRSs provide recommendations to groups, i.e., they take all individual group members’ preferences into account and satisfy them optimally with a sequence of items. Taking into consideration the fact that each group member has different behaviour with respect to other members in the group, we propose a genetic algorithm (GA) based multi-agent negotiation scheme for GRS (GA-MANS-GRS) where each agent acts on behalf of one group member. The GA-MANS-GRS is modelled as many one-to-one bilateral negotiation schemes with two phases. In the negotiation phase, we have applied GA to obtain the maximum utility offer for each user and generated the most appropriate ranking for each individual in the group. For the recommendation generation phase, again GA is employed to produce the list of ratings with that minimizes the sum of distances among the preferences of the group members. Finally, the results of computational experiments are presented that establish the superiority of our proposed model over baseline GRSs techniques.

     

一种面向多Agent协作和协商的消息机制

提出了一种基于概念模式匹配的消息传递机制,它使用了概念实例与概念模式相匹配的思想,实现了在具有多个Agent的系统中消息去向的灵活处理,从而很好地解决了多Agent之间的协作和协商问题。在引入多Agent协作与协商的概念后,讨论了这种机制的实现方法,并阐述了与这种机制实现相关的一些概念。     

MAPS: a system for multi-agent coordination

MAPS (Multi-Agent Planning System) is a system for multi-agent coordination that has been developed in the robot soccer domain. It uses potential fields to model the environment and provide directives for the robots. The key feature that enables successful coordination is the use of a shared world model. This paper describes MAPS and demonstrates it as a viable coordination system for use in multiple robot domains.     

Integrating multi-agent negotiation to resolve constraints in fulfilling supply chain orders

As the order fulfillment process (OFP) in supply chains shifts to outsourcing paradigm, the OFP performance relies on the coordination among supply chain partners to reach executable and effective plans. The coordination of OFP among supply chain partners can be viewed as a distributed constraint satisfaction problem (DCSP). This study adds the multi-agent negotiation mechanism to enhance the existing methods to solve the DCSP, and then evaluates the integrated system’s performance through experimentation on the OFP in the context of the metal industry. The experimental results show that the integrated system outperforms the existing distributed constraint satisfaction algorithms in various demand patterns.     

A multi-agent collaborative maintenance platform applying game theory negotiation strategies

Engineering asset management (EAM) is a broad discipline with distributed functions and services. When engineering assets are capital intensive, management requires specialized expertise for diagnosis, prognosis, maintenance and repairs. The current practice of EAM relies on self maintained experiential rules with coordinated collaboration and outsourcing for maintenance and repairs. In order to enhance the life long asset value and efficiency (from the stakeholder’s viewpoint) and after sales service quality (from the asset provider’s viewpoint), this research proposes a collaborative maintenance platform that integrates real time data collection with diagnostic and prognostic expertise. The collaborative system combines and delivers services among asset operation sites (the maintenance demanders), the service center (the intermediary coordinator), the system providers, the first tier maintenance collaborators, and the second and lower tier parts suppliers. Multi-agent system technology is used to integrate different systems and databases. Agents with autonomy and authority work to assist service providers and coordinate communications, negotiations, and maintenance decision support. Finally, game theory is used to design the decision models for strategic, tactical, and operational decision making during collaborative maintenance practices.     

MAPS: a language for multi-agent system design

Abstract: A language for multi-agent system design (MAPS) is presented and discussed in this paper. Any agent in MAPS is thought of as an expert system standing on its own. It can communicate through synchronous and asynchronous message sending. Dedicated behaviours are provided which specify how incoming messages are processed. Inter-agent cooperation is controlled via production rules. Two pre-defined agent classes are provided, which are given specific problem-solving roles: Knowledge Server (KS) agents are meant to maintain and transmit knowledge about problem-solving states, while Knowledge Processor (KP) agents are meant to process these elements in order to progress towards a solution. The duality between agent and resource modelling levels on the one hand and between KS and KP modelling styles on the other is shown to allow the specification of various control strategies. The environment is currently running on HP, SUN and DEC workstations.     

Cluster formation in a time-varying multi-agent system

We introduce time-varying parameters in a multi-agent clustering model and we derive necessary and sufficient conditions for the occurrence of clustering behavior with respect to a given cluster structure. For periodically varying parameters the clustering conditions may be formulated in a similar way as for the time-invariant model. The results require the individual weights assigned to the agents to be constant. For time-varying weights we illustrate with an example that the obtained results can no longer be applied.     

适用于工程设计的多智体协商冲突解决方法研究

讨论和分析了工程设计协作的特点和产生设计冲突的原因,根据分布式人工智能的思想,给出了适用于工程设计的多智体协商冲突解决模型和相关的冲突化解方法,其目的是通过工程设计过程中的冲突发现和冲突化解,提高设计效率,减少设计错误,降低成本。     

Development of a DRAM spots negotiation system using core hierarchy negotiation mechanism

The major factors of the DRAM negotiable transactions system in this research are the concession strategy and the negotiation mechanism of the sellers and buyers. In terms of negotiation mechanism, the core hierarchy model negotiation mechanism makes use of multistage sub-negotiates procedure to do multi-attribute negotiation for the sellers and buyers. The negotiable mechanism can further increase the successful rate of negotiation and the acceptable degree. In terms of concession strategy, research in the Liang and Doong, the concession strategy is constituted by times of negotiation, method of concession and times of persistence. There are less interactive concession for the sellers and buyers in the concession method of the negotiation in the related research. This research is to study the negotiable results that using various combinations of interactive concession strategy and the core hierarchy model negotiation mechanism. Be used as the reference of negotiation for the DRAM to trade on network.     

Knowledge-empowered automated negotiation system for e-Commerce

This paper focuses on knowledge empowered automated negotiation systems for buyer-centric multi-bilateral multi-attribute e-Procurement. We propose two knowledge empowered models, namely KERM and KACM. KERM is used for the buyer to determine a list of suppliers which are the best qualified candidates to negotiate with. The use of knowledge features largely in the model, which incorporates both the buyer’s and supplier’s profiles in evaluating a quote. Historical trade records of a supplier contribute to the supplier’s profile credit and therefore the rank of the supplier’s quote. KERM also allows the flexibility to assign appropriate weights, based on buyer’s interests, to each knowledge factor affecting the overall evaluation result of a quote. The resulted list of quotes of high rank is believed to produce satisfactory negotiation result for the buyer. KACM enables an automated concession process, while at the same time facilitates a flexible negotiation via the use of concept switch and tagged rules. Different from other negotiation models, KACM emphasizes the utilization of knowledge originated from the historical negotiation data in estimating and fine-tuning the negotiation parameters, for improving the performance of automated negotiation. Graph results show that our software prototype system makes significant improvement in the satisfaction level of negotiation results.

A queueing system with negative claims and a bunker for superseded claims in discrete time

We consider a discrete time single-line queueing system with independent geometric streams of regular and negative claims, infinite buffer, and geometric service. A negative claim pushes a regular claim out of the buffer queue and moves it to a bunker of infinite capacity. If the buffer is empty, a negative claim leaves the system without any effect. After servicing a claim, the system receives the next claim from the buffer, if it is not empty, or from the bunker. We obtain relations that allow computing stationary distributions for queues in the buffer and the bunker.     

Implicit: a multi-agent recommendation system for web search

For people with non-ordinary interests, it is hard to search for information on the Internet because search engines are impersonalized and are more focused on “average” individuals with “standard” preferences. In order to improve web search for a community of people with similar but specific interests, we propose to use the implicit knowledge contained in the search behavior of groups of users. We developed a multi-agent recommendation system called Implicit, which supports web search for groups or communities of people. In Implicit, agents observe behavior of their users to learn about the “culture” of the community with specific interests. They facilitate sharing of knowledge about relevant links within the community by means of recommendations. The agents also recommend contacts, i.e., who in the community is the right person to ask for a specific topic. Experimental evaluation shows that Implicit improves the quality of the web search in terms of precision and recall.     

Continuous auditing with a multi-agent system

Information technology has dramatically changed the way businesses and business information are managed. In electronic business, much of the information is in electronic format, and the resulting change in the auditing environment has forced audit professionals to audit electronic evidence. Moreover, the emergence of real time accounting reports has put increasing pressure on audit professionals to provide real-time auditing services, or continuous auditing, in which the time between the occurrence of events and the provision of an auditor's opinion is minimized to an acceptable level. This paper proposes an agent-based system for continuous auditing called the agent-based continuous audit model (ABCAM). The system can be implemented independently of the client's information system, is able to undertake automatic auditing in real time, and can easily adapt to changes in auditing requirements and information systems. Five scenarios are developed to illustrate the model.     

Towards a multi-agent system for regulated information exchange in crime investigations

This paper outlines a multi-agent architecture for regulated information exchange of crime investigation data between police forces. Interactions between police officers about information exchange are analysed as negotiation dialogues with embedded persuasion dialogues. An architecture is then proposed consisting of two agents, a requesting agent and a responding agent, and a communication language and protocol with which these agents can interact to promote optimal information exchange while respecting the law. Finally, dialogue policies are defined for the individual agents, specifying their behaviour within a negotiation. Essentially, when deciding to accept or reject an offer or to make a counteroffer, an agent first determines whether it is obligatory or permitted to perform the actions specified in the offer. If permitted but not obligatory, the agent next determines whether it is in his interest to accept the offer.     

Engineering a multi-agent system in Jason and CArtAgO

This paper presents the overall strategy utilized by Jason-DTU to achieve a shared second place in the annual Multi-Agent Programming Contest. It provides an overview of the implementation details considering perception, task handling, agent logic and more. The paper analyzes the team’s results in each match, and is summarized by evaluating strengths and weaknesses of the proposed multi-agent system.     

egoisst: a negotiation support system for electronic business-to-business negotiations in e-commerce

Electronic negotiations can range from simple offer exchanges to complex communicative acts concerning packages of products and services. In contrast to dominant approaches aiming at automating the negotiation process (e.g. auction models), we introduce the notion of negotiation support for human negotiators conducting complex electronic negotiations. The negotiation support system egoisst for business-to-business electronic commerce is presented that is based on theories of communication and information systems and that combines communication and document management. egoisst has been successfully validated for e-negotiations in the construction industry.     

On temporal characteristics in an exponential queueing system with negative claims and a bunker for ousted claims

We consider a queueing system with Poisson input streams of positive and negative claims, an infinite collector, and exponential service. A negative claim ousts a positive claim out of the collector queue and moves it to a bunker of unbounded capacity. If the collector is empty then a negative claim leaves the system with no influence on it. After a claim is serviced, the device receives the next claim from the collector or, if the collector is empty, from the bunker. For different combinations of FIFO and LIFO orders of choosing a claim for service from the collector’s queue, choosing a claim for service from the bunker’s queue, and ousting claims from the collector to the bunker, we obtain formulas for computing the stationary waiting time distribution for a claim to begin service and other temporal characteristics.     

A Q-learning-based multi-agent system for data classification

In this paper, a multi-agent classifier system with Q-learning is proposed for tackling data classification problems. A trust measurement using a combination of Q-learning and Bayesian formalism is formulated. Specifically, a number of learning agents comprising hybrid neural networks with Q-learning, which we have formulated in our previous work, are devised to form the proposed Q-learning Multi-Agent Classifier System (QMACS). The time complexity of QMACS is analyzed using the big O-notation method. In addition, a number of benchmark problems are employed to evaluate the effectiveness of QMACS, which include small and large data sets with and without noise. To analyze the QMACS performance statistically, the bootstrap method with 95% confidence interval is used. The results from QMACS are compared with those from its constituents and other models reported in the literature. The outcome indicates the effectiveness of QMACS in combining the predictions from its learning agents to improve the overall classification performance.