A Multi-Agent Coalition Formation Method Based on Preference Models

Coordination is one of the fundamental research issues in distributed artificial intelligence and multi-agent systems. Current multi-agent coalition formation methods present two major problems. First, some of these methods can be applied only to cooperative multi-agent systems; second, the algorithms proposed may fail in the formation of the coalitions. This article proposes two methods for problems of agent coalition formation in cooperative and non-cooperative multi-agent systems. These methods are based on agent preference models and on preference aggregation using the Choquet integral. Our first method based on the evolutionary system design (ESD) methodology proposed by M. Shakun is required for competitive multi-agent systems, whereas our second method is powerful for cooperative multi-agent systems where the agents can exchange their information.     

Coalition Formation Procedures: The Impact of Issue Saliences and Consensus Estimation

Coalition formation procedures incorporate two properties that are not often found in other coalition formation models: the choice between different formation paths and constrained consensus positions. However, there are two aspects of coalition formation procedures that are often overlooked: issue saliences and consensus estimation. Issue saliences are a measure of the importance that parties can attribute to issue dimensions. Initially, we employ the classical application to implement issue saliences. The classical application combines the Euclidean distance with the center of gravity as a consensus estimate. Secondly, we introduce a consistent distance application where the coalition consensus position is determined by minimizing the sum of salience-weighted Euclidean distances. The impact of these aspects is examined with the help of both numerical and empirical applications. The results indicate that both the consensus estimation method and the inclusion of issue saliences do not only have an impact on the estimated consensus position. They also determine the individual parties’ preferences towards the potential coalition formation procedures.     

Equilibrium coalition structures in the presence of foreign direct investment

This paper incorporates foreign direct investment (FDI) into the examination of trading bloc formation with endogenously determined coalition structures. In so doing, we build a three-country model, in which firms serve foreign markets either by exporting or undertaking FDI, and consider a coalition formation game with the Coalition Proof Nash Equilibrium as an equilibrium concept. We find that the equilibrium coalition structure varies upon firms characterization before and after the formation of a trading bloc. As in the literature, when all firms are exporters in the pre- and post-formation, bilateralism can be an equilibrium outcome. However, when trade barriers are not so high as to be trade-prohibitive and the environment is favorable to multinational activities in the pre- or post-formation, only global free trade will prevail as an equilibrium coalition structure.     

Meta-level Coordination for Solving Distributed Negotiation Chains in Semi-cooperative Multi-agent Systems

A negotiation chain is formed when multiple related negotiations are spread over multiple agents. In order to appropriately order and structure the negotiations occurring in the chain so as to optimize the expected utility, we present an extension to a single-agent concurrent negotiation framework. This work is aimed at semi-cooperative multi-agent systems, where each agent has its own goals and works to maximize its local utility; however, the performance of each individual agent is tightly related to other agents’ cooperation and the system’s overall performance. We introduce a pre-negotiation phase that allows agents to transfer meta-level information. Using this information, the agent can improve the accuracy of its local model about how other agents would react to the negotiations. This more accurate model helps the agent in choosing a better negotiation solution for a distributed negotiation chain problem. The agent can also use this information to allocate appropriate time for each negotiation, hence to find a good ordering of all related negotiations. The experimental data show that these mechanisms improve the agents’ and the system’s overall performance significantly.     

Negotiating a Stable Government: An Application of Bargaining Theory to a Coalition Formation Model

In this paper, we apply bargaining theory to a certain model of coalition formation. The notions of a feasible government and a stable government are central in the model considered. By a government, we mean a pair consisting of a majority coalition and a policy supported by this coalition. The aim of this paper is to establish which stable government should be created if more than one stable government exists or, in case there is no stable one, which feasible government should be formed if more than one feasible government exists. Several bargaining procedures leading to the choice of one stable (or feasible) government are proposed. We define bargaining games in which only parties belonging to at least one stable (or feasible) government bargain over the creation of a government. We consider different bargaining costs. We investigate subgame perfect equilibria of the bargaining games defined. It turns out that the prospects of a party depend on the procedure applied, and on the bargaining costs assumed. We also apply the coalition formation model to the Polish Parliament after the 2001 elections and apply the different bargaining games for the creation of a government to this example.     

Argumentation-Based Qualitative Preference Modelling with Incomplete and Uncertain Information

This paper presents an argumentation-based framework for the modelling of, and automated reasoning about multi-attribute preferences of a qualitative nature. The framework presents preferences according to the lexicographic ordering that is well-understood by humans. Preferences are derived in part from knowledge. Knowledge, however, may be incomplete or uncertain. The main contribution of the paper is that it shows how to reason about preferences when only incomplete or uncertain information is available. We propose a strategy that allows reasoning with incomplete information and discuss a number of strategies to handle uncertain information. It is shown how to extend the basic framework for modelling preferences to incorporate these strategies.     

The G20 at the Cancun Ministerial: Developing Countries and Their Evolving Coalitions in the WTO

In large measure, the voice that developing countries were able to exercise in Cancun was a result of their effective coalition formation. In this paper we present a brief overview of the various coalitions that played an important role at Cancun. The greater part of this paper focuses on one among these various coalitions: the G20 on agriculture. The G20 presents an especially fascinating case of a coalition that combined a great diversity of members and apparently incompatible interests. All theoretical reasoning and historical precedent predicted that the group would collapse in the endgame. And yet the group survived. We investigate the sources of the unity of this group and trace them to a process of learning that allowed the group to acquire certain structural features and develop strategies that helped to cement it further. While our central dependent variable is the cohesion of the G20, we also address the derivative question of the costs and benefits of maintaining such coalitions. The Cancun coalitions give us an excellent case of coalitions that managed to retain their cohesion, but also ended up with a situation of no agreement rather than a fulfilment of even some of their demands. We examine some of the causes behind the impasse in the negotiation process and suggest ways in which future outcomes could be improved.     

Class alliances and conflict: An explanation of political transitions in 19th‐century Europe

I develop a model analyzing common interests and conflict among four classes—capitalists, workers, landlords, and peasants in 19th‐century Europe—and show that strong class position, based on a high degree of organization and solidarity, may actually be detrimental to the economic and political advantage of that class. This occurs when a strong class is excluded from a major class coalition via coalition formation processes. The reason is that the weak class, if they enjoy bargaining power over even weaker classes within a coalition, may not want to form a coalition with the strong class. I apply the main results to coalition formation and political transitions in 19th‐century European society.     

Using Bids,Arguments and Preferences in Sensitive Multi-unit Assignments: A p-Equitable Process and a Course Allocation Case Study

Bonus distribution in enterprises or course allocation at universities are examples of sensitive multi-unit assignment problems, where a set of resources is to be allocated among a set of agents having multi-unit demands. Automatic processes exist, based on quantitative information, for example bids or preference ranking, or even on lotteries. In sensitive cases, however, decisions are taken by persons also using qualitative information. At present, no multi-unit assignment system supports both quantitative and qualitative information. In this paper, we propose MUAP-LIS, an interactive process for multi-assignment problems where, in addition to bids and preferences, agents can give arguments to motivate their choices. Bids are used to automatically make pre-assignments, qualitative arguments and preferences help decision makers break ties in a founded way. A group decision support system, based on Logical Information Systems, allows decision makers to handle bids, arguments and preferences in a unified interface. We say that a process is p-equitable for a property p if all agents satisfying p are treated equally. We formally demonstrate that MUAP-LIS is p-equitable for a number of properties on bids, arguments and preferences. It is also Pareto-efficient and Gale–Shapley-stable with respect to bids. A successful course allocation case study is reported. It spans over two university years. The decision makers were confident about the process and the resulting assignment. Furthermore, the students, even the ones who did not get all their wishes, found the process to be equitable.     

Optimization in electronic markets: examples in combinatorial auctions

Combinatorial auctions provide an important tool for mechanism design in multi-agent systems. When implemented they require to solve combinatorial optimization problems such as set packing and partitioning problems. We present in this paper an analysis of the complexity of the problem to assign bids to bidders in combinatorial auctions. We show that the case of identical assets can be solved in polynomial time. The case of non-identical assets is in its general version NP-hard. Extra structure, like a complete ordering of assets, or mild side conditions make the problem solvable. Finally, we present an algorithm to solve small and medium sized instances in a limited time using standard software.     

Coalition formation in international monetary policy games

This paper analyses coalition formation in monetary policy coordination games between n countries. We show that some but not all countries may join if the decision to be a member of the coalition is incentive-compatible for the individual country. Positive spillovers of the coalition formation process and the resulting free-rider problem limit the stable coalition size: since the coalition members are bound by the union’s discipline, an outsider can successfully export inflation without fearing that the insiders will try to do the same. These ‘gains from staying out’ arise even in the case of symmetric shocks.     

Indirect information exchange

The goal of this paper is to show that when agents can engage in second-hand trading of information, some agents may not be able to exchange their information with other agents. With three agents, such foreclosure is possible only when agents can refuse to exchange information. With four or more players, foreclosure is possible even when players cannot abstain from or refuse to trade, but it does not constitute a subgame perfect equilibrium. Players can avoid being foreclosed by choosing what to trade and with whom. The results can be applied to formal and informal information sharing, file sharing systems and peer-to-peer networks.     

An improved collaborative filtering approach for predicting cross-category purchases based on binary market basket data

Retail managers have been interested in learning about cross-category purchase behavior of their customers for a fairly long time. More recently, the task of inferring cross-category relationship patterns among retail assortments is gaining attraction due to its promotional potential within recommender systems used in online environments. Collaborative filtering algorithms are frequently used in such settings for the prediction of choices, preferences and/or ratings of online users. This paper investigates the suitability of such methods for situations when only binary pick-any customer information (i.e., choice/non-choice of items, such as shopping basket data) is available. We present an extension of collaborative filtering algorithms for such data situations and apply it to a real-world retail transaction dataset. The new method is benchmarked against more conventional algorithms and can be shown to deliver superior results in terms of predictive accuracy.     

Islamic economics and a third fundamental theorem of welfare economics

The discipline of economics started as a moral science but became detached from moral concerns over time to emulate natural science and to adopt positivism. Consequently, mainstream economics assumes people to be sordidly selfish. The teachings of Islam, however, promote social preferences where individuals should be other‐regarding and have preferences over social outcomes. This paper replaces the selfish agent with a social agent and presents the results in a theorem referred to as the third fundamental theorem of welfare economics (TFTWE). The TFTWE states that “when the selfish agents are replaced with the social agents, market outcomes are Pareto optimal, equitable, and unique”. This is an important result which has widespread implications. We show that the TFTWE holds under conditions where the first two fundamental theorems of welfare economics fail and that a Walrasian equilibrium is more likely to exist when selfish preferences are non‐convex. Unlike the popular convention, there is no equity‐efficiency trade‐off. In fact we point to the possibility of reversal in equity‐efficiency trade‐off.     

Coalition Formation: The Role of Procedure and Policy Flexibility

A spatial model of coalition formation is used together with data from Dutch elections and theoretical instances to study different procedures of coalition formation. The model shows that procedure plays an important role in reaching a coalition agreement and that political parties do not necessarily benefit from being a first-mover. Moreover, it is shown that a decrease in a party’s flexibility can be (dis)advantageous in coalition negotiations. Furthermore, certain power sharing tactics appear not always to lead to an agreement that is in a party’s advantage. The main message put forward is that the procedure of forming a coalition plays a more important role than is usually acknowledged in literature and practice.     

A Multi-Agent Model for Overlapping Negotiations

In the last few years, research on multi-agent systems has addressed different aspects of intelligent negotiations using methods developed in different domains including game theory, decision theory, and economic models. The research proposed in Andersson and Sandholm (1999), Sandholm (1993), Sandholm and Lesser (1995) and Aknine et?al. (2004), Aknine (2002) are significant examples. However, only some of this work focuses on problems related to complex negotiations, particularly those concerning new generation applications. This new research raises fundamental difficulties we have encountered, especially in overlapping negotiations and combined negotiations. This article is interested essentially in overlapping negotiations, which include several agent roles in a same negotiation. One or more agents may play each of these roles. This work shows that the high-level negotiation models are necessary in order to control the execution of overlapping negotiation processes, since, in these negotiations we are facing both classical problems of multi-agent negotiations based on two agents’ roles and the problems concerned with the interdependence of these negotiations. Synchronization of these different processes is thus necessary because of the multiplicity of the roles. Thus, this paper presents a formalized negotiation model, which deals with this problem. It gives a theoretical analysis of the suggested model and discusses the results of the experimental evaluation. To perform this evaluation, we use the application of intelligent service agencies on the Internet.     

When Shelf-Based Scarcity Impacts Consumer Preferences

Shelf-based scarcity in the form of relative stocking level depletion significantly affects consumer preferences (van Herpen et al. 2009). While both popularity and quality inferences are induced by stocking-level depletion, this paper demonstrates that popularity (rather than quality) inferences are the primary driver of the effect. Shelf-based scarcity is shown to have an impact in the presence of other signals, such as brand names and quality ratings, which indicates its effect may be widespread. However, its effect on preferences is greatly decreased by the presence of strong prior preferences and/or price promotions, indicating that shelf-based scarcity is likely to affect choices when consumers lack strong prior preferences and under conditions where price promotions are either not present or are similar across alternatives.     

Retailers’ coalition and quantity discounts under demand uncertainty

Coalition of retailers is a nowadays phenomenon in retailing channels (RCs) that makes it possible for the retailers to enhance their business performance and respond to the consumers' needs more effectively. This paper discusses the retailers' coalition advantages and challenges in a two-echelon retailing channel consist of one wholesaler and two non-competing retailers who serve the consumers. The model is developed under a classic newsvendor problem where an all-unit quantity discount is offered by the wholesaler. Our investigations are conducted for both homogeneous/heterogeneous retailers under three different scenarios. In the developed models, the retailers start to make a coalition and make a joint order in order to gain more from the offered discount by the wholesaler, and then using a heuristic procedure they allocate the received order to serve their individual markets. The main objective of this study is to analyze and resolve challenges of the retailers' coalition when a quantity discount is offered in order to make them capable to meet consumers' needs in the best possible manner. Three models are developed: (1) no-coalition decentralized scheme, (2) decentralized coalition where the retailers coalesce, and (3) centralized coalition where not only the retailers but the wholesaler participate in the coalition scheme. Our paper contributes to the RC management literature by analyzing the benefits of the retail coalition in taking most advantage from an offered quantity discount. Results demonstrate that the optimal configuration of the retailers' coalition increases profits of both the retailers and the RC in comparison with the no-coalition scenario. Our findings help procurement managers to rethink their ordering policy toward forming coalition to gain more profits and enhance their service level to meet consumers’ needs.     

Qualitative One-to-Many Multi-Issue Negotiation: Approximating the QVA

When there is one buyer interested in obtaining a service from one of a set of sellers, multi-attribute or multi-issue auctions can ensure an allocation that is efficient. Even when there is no transferable utility (e.g., money), a recent qualitative version of the Vickrey auction may be used, the QVA, to obtain a Pareto-efficient outcome where the best seller wins. However, auctions generally require that the preferences of at least one party participating in the auction are publicly known, while often making this information public is costly, undesirable, or even impossible. It would therefore be useful to have a method that does not impose such a requirement, but is still able to approximate the outcome of such an auction. The main question addressed here is whether the Pareto-efficient best-seller outcome in multi-issue settings without transferable utility (such as determined by the QVA) can be reasonably approximated by multi-bilateral closed negotiation between a buyer and multiple sellers. In these closed negotiations parties do not reveal their preferences explicitly, but make alternating offers. The main idea is to have multiple rounds of such negotiations. We study three different variants of such a protocol: one that restricts the set of allowed offers for both the buyer and the seller, one where the winning offer is announced after every round, and one where the sellers are only told whether they have won or not after every round. It is shown experimentally that this protocol enables agents that can learn preferences to obtain agreements that approximate the Pareto-efficient best-seller outcome as defined by the auction mechanism. We also show that the strategy that exploits such a learning capability in negotiation is robust against and dominates a Zero Intelligence strategy. It thus follows that the requirement to publicly announce preferences can be removed when negotiating parties are equipped with the proper learning capabilities and negotiate using the proposed multi-round multi-bilateral negotiation protocol.