Alon Grubshtein

Methods for Computing Trust and Reputation While Preserving Privacy

Trust and Reputation systems in distributed environments attain widespread interest as online communities are becoming an inherent part of the daily routine of Internet users. Trust-based models enable safer operation within communities to which information exchange and peer to peer interaction are centric. Several models for trust based reputation have been suggested recently, among them the Knots model [5]. In these models, the subjective reputation of a member is computed using information provided by a set of members trusted by the latter. The present paper discusses the computation of reputation in such models, while preserving members private information. Three different schemes for the private computation of reputation are presented, and the advantages and disadvantages in terms of privacy and communication overhead are analyzed.

Asymmetric distributed constraint optimization problems

Distributed Constraint Optimization (DCOP) is a powerful framework for representing and solving distributed combinatorial problems, where the variables of the problem are owned by different agents. Many multi-agent problems include constraints that produce different gains (or costs) for the participating agents. Asymmetric gains of constrained agents cannot be naturally represented by the standard DCOP model. n nThe present paper proposes a general framework for Asymmetric DCOPs (ADCOPs). In ADCOPs different agents may have different valuations for constraints that they are involved in. The new framework bridges the gap between multi-agent problems which tend to have asymmetric structure and the standard symmetric DCOP model. The benefits of the proposed model over previous attempts to generalize the DCOP model are discussed and evaluated. n nInnovative algorithms that apply to the special properties of the proposed ADCOP model are presented in detail. These include complete algorithms that have a substantial advantage in terms of runtime and network load over existing algorithms (for standard DCOPs) which use alternative representations. Moreover, standard incomplete algorithms (i.e., local search algorithms) are inapplicable to the existing DCOP representations of asymmetric constraints and when they are applied to the new ADCOP framework they often fail to converge to a local optimum and yield poor results. The local search algorithms proposed in the present paper converge to high quality solutions. The experimental evidence that is presented reveals that the proposed local search algorithms for ADCOPs achieve high quality solutions while preserving a high level of privacy.

Concurrent forward bounding for distributed constraint optimization problems

A distributed search algorithm for solving Distributed Constraints Optimization Problems (DCOPs) is presented. The new algorithm scans the search space by using multiple search processes (SPs) that run on all agents concurrently. SPs search in non-intersecting parts of the global search space and perform Branch & Bound search. Each search process (SP) uses the mechanism of forward bounding (FB) to prune efficiently its part of the global search space. The Concurrent Forward-Bounding (ConcFB) algorithm enables all SPs to share their upper bound across all parts of the global search space. The number of concurrent SPs is controlled dynamically by the ConcFB algorithm, by performing dynamic splitting. Within each SP a dynamic variable ordering is employed in order to help control the balance of computational load among all agents and across different SPs. The ConcFB algorithm is evaluated experimentally and compared to all state of the art DCOP algorithms. The number of Non-Concurrent Logical Operations, Non-Concurrent Steps, the total number of messages sent and CPU time are used as performance metrics. The evaluation procedure considers different DCOP problem types with a varying number of agents and different constraint graphs. As problems become larger and denser, ConcFB is shown to outperform all other evaluated algorithms by 2-3 orders of magnitude in all performance measures. Further evaluations comparing different variants of ConcFB provide important insights into the working of the algorithm and reveals the contribution of its different components.

Hybrid search for minimal perturbation in Dynamic CSPs

It is often the case that after a scheduling problem has been solved some small changes occur that make the solution of the original problem not valid. Solving the new problem from scratch can result in a schedule that is very different from the original schedule. In applications such as a university course timetable or flight scheduling, one would be interested in a solution that requires minimal changes for the users. The present paper considers the minimal perturbation problem. It is motivated by scenarios in which a Constraint Satisfaction Problem (CSP) is subject to changes. In particular, the case where some of the constraints are changed after a solution was obtained. The goal is to find a solution to the changed problem that is as similar as possible (e.g. includes minimal perturbations) to the previous solution. Previous studies proposed a formal model for this problemxa0(Barták etxa0al. 2004), a best first search algorithmxa0(Ross et al. 2000), complexity boundsxa0(Hebrard et al. 2005), and branch and bound based algorithmsxa0(Barták etxa0al. 2004; Hebrard et al. 2005). The present paper proposes a new approach for solving the minimal perturbation problem. The proposed method interleaves constraint optimization and constraint satisfaction techniques. Our experimental results demonstrate the advantage of the proposed algorithm over former algorithms. Experiments were performed both on random CSPs and on random instances of the Meeting Scheduling Problem.

A Distributed Cooperative Approach for Optimizing a Family of Network Games

The present study considers a distributed cooperative approach for network problems where agents have personal preferences over outcomes. Such problems can be described by Asymmetric Constraints where the joint action of agents yields different gains to each participant Grubshtein et al. (2010). The proposed method constructs and solves an Asymmetric Distributed Constraints Optimization Problem whose solutions guarantee a minimal gain for each agent, which is at least as high as the agents’ Bayesian equilibrium gain. The paper focuses on a special class of Network Games and proves that the proposed method produces optimal results in terms of the number of agents whose gain improves over their equilibrium gain and that the resulting solutions are Pareto Efficient. Extensive empirical evaluation of the studied network problem shows that the number of improving agents is not negligible and that under some configurations up to 70% of the agents improve their gain while none of the agents receive a payoff lower than their equilibrium gain.

AgentZero: A Framework for Simulating and Evaluating Multi-agent Algorithms

Applications of multi-agent system (MAS) are versatile. In this chapter we focus on a specific application domain—agent-oriented programming for distributed constraint reasoning (DCR ). The field of DCR deals with constraints -based problems that are distributed among multiple agents. The agents need to arrive at an optimal solution to the global combinatorial problem, and in order to do so, they run a distributed search algorithm . Another important aspect of MAS software development is MAS simulation . In this regard, this chapter introduces a new agent-based research tool for designing and testing DCR algorithms. The new tool—AgentZero—is specifically designed for the specification, implementation, and evaluation of DCR search algorithms. AgentZero provides full support to researchers of distributed constraints algorithms in the form of an extensive agent-based environment for algorithmic research that includes a distributed run-time environment, built-in performance measures that are automatically used by all algorithms, and visualization tools that help design and understand the behavior of complex distributed search algorithm s. The API of the AgentZero simulator is described in detail and important architectural decisions that enable analysis and smooth implementation of a variety of algorithms are explained and described. In the context of AOSE, this chapter exemplifies two aspects: agent-based simulation environment and tools, and a variety of development and runtime aids for agent-based systems.

Finding a Nash Equilibrium by Asynchronous Backtracking

Graphical Games are a succinct representation of multi agent interactions in which each participant interacts with a limited number of other agents. The model resembles Distributed Constraint Optimization Problems (DCOPs) including agents, variables, and values (strategies). However, unlike distributed constraints, local interactions of Graphical Games take the form of small strategic games and the agents are expected to seek a Nash Equilibrium rather than a cooperative minimal cost joint assignment. n nThe present paper models graphical games as a Distributed Constraint Satisfaction Problem with unique k-ary constraints in which each agent is only aware of its part in the constraint. A proof that a satisfying solution to the resulting problem is an e-Nash equilibrium is provided and an Asynchronous Backtracking algorithm is proposed for solving this distributed problem. The algorithms completeness is proved and its performance is evaluated.

Cost of Cooperation for Scheduling Meetings

Scheduling meetings among agents can be represented as a game - the Meetings Scheduling Game (MSG). In its simplest form, the two-person MSG is shown to have a price of anarchy (PoA) which is bounded by 0.5. The paper defines the Cost of Cooperation (CoC) for meetings scheduling games, with respect to different global objective functions. For an “egalitarian” objective, that maximizes the minimal gain among all participating agents, the CoC is non positive for all agents. This makes the MSG a cooperation game. The concepts are defined and examples are given within the context of the MSG. A game may be revised by adding a mediator (or with a slight change of its mechanism) so that it behaves as a cooperation game. Thus, rational participants can cooperate (by taking part in a distributed optimization protocol) and receive a payoff which will be at least as high as the worst gain expected by a game theoretic equilibrium point.

Partial cooperation in multi-agent local search

Multi-agent systems usually address one of two forms of interaction. One has completely competitive agents that act selfishly, each maximizing its own gain from the interaction. Auctions and voting scenarios usually assume such agents and follow game theoretic results. The other form of interaction has multiple agents that cooperatively search for some global goal, such as an optimal time slot allocation for all landing aircrafts in an airport. n nThe present paper proposes a paradigm for multiple agents solving a distributed problem using local search algorithms and acting in a partially cooperative manner. That is, agents with different preferences search for a minimal cost solution to an Asymmetric Distributed Constraints Optimization Problem (ADCOP), while keeping a limited form of self interest. n nTwo approaches for using local search in the partial cooperative paradigm are proposed. The first, modifies the anytime mechanism introduced by Zivan [19] so that agents can eliminate solutions which do not satisfy their cooperation thresholds. The second proposes a new local search algorithm that explores only valid solutions. n nThe performance of two innovative algorithms implementing these two approaches, are compared with state of the art local search algorithms on three different setups. When personal constraints are strict, the proposed algorithms have a large advantage over existing algorithms. We provide insights to the success of existing algorithms within the anytime framework when constraints are loose.