Yingqian Zhang

Distributed task allocation in social networks

This paper proposes a new variant of the task allocation problem, where the agents are connected in a social network and tasks arrive at the agents distributed over the network. We show that the complexity of this problem remains NP-hard. Moreover, it is not approximable within some factor. We develop an algorithm based on the contract-net protocol. Our algorithm is completely distributed, and it assumes that agents have only local knowledge about tasks and resources. We conduct a set of experiments to evaluate the performance and scalability of the proposed algorithm in terms of solution quality and computation time. Three different types of networks, namely small-world, random and scale-free networks, are used to represent various social relationships among agents in realistic applications. The results demonstrate that our algorithm works well and that it scales well to large-scale applications.

On the Complexity of Efficiency and Envy-Freeness in Fair Division of Indivisible Goods with Additive Preferences

We study the problem of allocating a set of indivisible goods to a set of agents having additive preferences. We introduce two new important complexity results concerning efficiency and fairness in resource allocation problems: we prove that the problem of deciding whether a given allocation is Pareto-optimal is coNP-complete, and that the problem of deciding whether there is a Pareto-efficient and envy-free allocation is

Impact: A Multi-Agent Framework with Declarative Semantics

\Sigma_2^p

Monitoring agents using declarative planning

-complete.

Computing the fault tolerance of multi-agent deployment

The IMPACT project (http://www.cs.umd.edu/projects/impact) aims at developing a powerful multi-agent system platform, which (1) is able to deal with heterogenous and distributed data, (2) can be realised on top of arbitrary legacy code, (3) is built on a clear foundational basis, and (4) scales up for realistic applications. We will describe its main features and several extensions of the language that have been investigated (and partially implemented).

Coordination by design and the price of autonomy

In this paper we consider the following problem: Given a particular description of a multi-agent system (MAS), is it implemented properly? We assume that we are given (possibly incomplete) information about the system and aim at refuting its proper implementation. In our approach, agent collaboration is described as an action theory. Action sequences reaching the collaboration goal are computed by a planner, whose compliance with the actual MAS behaviour allows to detect possible collaboration failures. The approach can be fruitfully applied to aid in offline testing of a MAS implementation, as well as in online monitoring.

Two-agent scheduling on a single parallel-batching machine with equal processing time and non-identical job sizes

A deployment of a multi-agent system on a network refers to the placement of one or more copies of each agent on network hosts, in such a manner that the memory constraints of each node are satisfied. Finding the deployment that is most likely to tolerate faults (i.e. have at least one copy of each agent functioning and in communication with other agents) is a challenge. In this paper, we address the problem of finding the probability of survival of a deployment (i.e. the probability that a deployment will tolerate faults), under the assumption that node failures are independent. We show that the problem of computing the survival probability of a deployment is at least NP-hard. Moreover, it is hard to approximate. We produce two algorithms to accurately compute the probability of survival of a deployment-these algorithms are expectedly exponential. We also produce five heuristic algorithms to estimate survival probabilities-these algorithms work in acceptable time frames. We report on a detailed set of experiments to determine the conditions under which some of these algorithms perform better than the others.

Fair task allocation in transportation

We consider a multi-agent planning problem as a set of activities that has to be planned by several autonomous agents. In general, due to the possible dependencies between the agents’ activities or interactions during execution of those activities, allowing agents to plan individually may lead to a very inefficient or even infeasible solution to the multi-agent planning problem. This is exactly where plan coordination methods come into play. In this paper, we aim at the development of coordination by design techniques that (i) let each agent construct its plan completely independent of the others while (ii) guaranteeing that the joint combination of their plans always is coordinated. The contribution of this paper is twofold. Firstly, instead of focusing only on the feasibility of the resulting plans, we will investigate the additional costs incurred by the coordination by design method, that means, we propose to take into account the price of autonomy: the ratio of the costs of a solution obtained by coordinating selfish agents versus the costs of an optimal solution. Secondly, we will point out that in general there exist at least two ways to achieve coordination by design: one called concurrent decomposition and the other sequential decomposition. We will briefly discuss the applicability of these two methods, and then illustrate them with two specific coordination problems: coordinating tasks and coordinating resource usage. We also investigate some aspects of the price of autonomy of these two coordination methods.

A Tabu Search Algorithm for application placement in computer clustering

We schedule the jobs from two agents on a single parallel-batching machine with equal processing time and non-identical job sizes. The objective is to minimize the makespan of the first agent subject to an upper bound on the makespan of the other agent. We show that there is no polynomial-time approximation algorithm for solving this problem with a finite worst-case ratio, unless P=NP. Then, we propose an effective algorithm LB to obtain a lower bound of the optimal solution, and two algorithms, namely, reserved-space heuristic (RSH) and dynamic-mix heuristic (DMH), to solve the two-agent scheduling problem. Finally, we evaluate the performance of the proposed algorithms with a set of computational experiments. The results show that Algorithm LB works well and tends to perform better with the increase of the number of jobs. Furthermore, our results demonstrate that RSH and DMH work well on different cases. Specifically, when the optimal makespan on the first agent exceeds the upper bound of the makespan of the other agent, RSH outperforms or equals DMH, otherwise DMH is not less favorable than RSH.

Distributed algorithms for dynamic survivability of multiagent systems

Task allocation problems have traditionally focused on cost optimization. However, more and more attention is being given to cases in which cost should not always be the sole or major consideration. In this paper we study a fair task allocation problem in transportation where an optimal allocation not only has low cost but more importantly, it distributes tasks as even as possible among heterogeneous participants who have different capacities and costs to execute tasks. To tackle this fair minimum cost allocation problem we analyze and solve it in two parts using two novel polynomial-time algorithms. We show that despite the new fairness criterion, the proposed algorithms can solve the fair minimum cost allocation problem optimally in polynomial-time. In addition, we conduct an extensive set of experiments to investigate the trade-off between cost minimization and fairness. Our experimental results demonstrate the benefit of factoring fairness into task allocation. Among the majority of test instances, fairness comes with a very small price in terms of cost.