Abstract — Two forms of cooperation in distributed problem solving are considered: task-sharing and result-sharing. In the former, nodes assist each other by sharing the computational load for the execution of subtasks of the overall problem. In the latter, nodes assist each other by sharing partial results which are based on somewhat different perspectives on the overall problem. Different perspectives arise because the nodes use different knowledge sources (KS’s) (e.g., syntax versus acoustics in the case of a speech-understanding system) or different data (e.g., data that is sensed at different locations in the case of a distributed sensing system). Particular attention is given to control and to internode communication for the two forms of cooperation. For each, the basic methodology is presented and systems in which it has been used are described. The two forms are then compared and the types of applications for which they are suitable are considered. I. DISTRIBUTED PROBLEM SOLVING

Introduction Cooperative Distributed Problem-Solving (CDPS) studies how a loosely-coupled network of problem solvers can work together to solve problems that are beyond their individual capabilities. Each problem-solving node in the network is capable of sophisticated problem solving and can work independently, but the problems faced by the nodes cannot be completed without cooperation. Cooperation is necessary because no single node has sufficient expertise, resources, and information to solve a problem, and different nodes might have expertise for solving different parts of the problem. For example, if the problem is to design a house, one node might have expertise on the strength of structural materials, another on the space requirements for different types of rooms, another on plumbing, another on electrical wiring, and so on. Different nodes might have different resources: some might be very fast at computation, others might have connections that speed communication, whil

A new approach for structuring distributed processing systems, called functionally accurate, cooperative (FA/C), is proposed. The approach differs from conventional ones in its emphasis on handling distribution-caused uncertainty and errors as an integral part of the network problem-solving process. In this approach nodes cooperatively problem solve by exchanging partial tentative results (at various levels of abstraction) within the context of common goals. The approach is especially suited to applications in which the data necessary to achieve a solution cannot be partitioned in such a way that a node can complete a task without seeing the intermediate state of task processing at other nodes. Much of the inspiration for the FA/C approach comes from the mechanisms used in knowledge-based artificial intelligence (AI) systems for resolving uncertainty caused by noisy input data and the use of approximate knowledge. The appropriateness of the FA/C approach is explored in three application domains: distributed interpretation, distributed network traffic-light control, and distributed planning. Additionally, the relationship between the approach and the structure of management organizations is developed. Finally, a number of current research directions necessary to more fully develop the FA/C approach are outlined. These research directions include distributed search, the integration of implicit and explicit forms of control, and distributed planning and organizational self-design. I.

This paper was presented at IJCAI-83. Distributed problem-solving networks provide an interesting application area for meta-level control through the use of organizational structuring. We describe a decentralized approach to network coordination that relies on each node making sophisticated local decisions that balance its own perceptions of appropriate problem-solving activity with activities deemed important by other nodes. Each node is guided by a high-level strategic plan for cooperation among the nodes in the network. The high-level strategic plan, which is a form of meta-level control, is represented as a network organizational structure that specifies in a general way the information and control relationships among the nodes. An implementation of these ideas is briefly described along with the results of preliminary experiments with various network problem-solving strategies specified via organizational structuring. In addition to its application to Distributed Artificial Intelligence, this research has implications for organizing and controlling complex knowledge-based systems that involve semi-autonomous problem solving agents. 1

We are implementing a control system for a discrete manufacturing environment that partitions tasks using a negotiation protocol like the contract net described by Smith and Davis [24,25,26,3]. The application domain differs in interesting ways from those to which contract nets have previously been applied. This report

Research in Distributed Artificial Intelligence is concerned with how automated agents can be designed to interact effectively. One important capability that could aid inter-agent cooperation would be that of negotiation: agents could be built that are able to communicate their respective desires and compromise to reach mutually beneficial agreements. This work uses the language of game theory to analyze negotiation among automated agents in cooperative domains. However, while game theory generally deals with negotiation in continuous domains and among agents with full information, this research considers discrete domains and the case where agents have only partial information, assumptions of greater interest for artificial intelligence. A novel, stable, negotiation protocol is introduced for the case of agents who are able to share a discrete set of tasks with one another. The case of agents who may lie to one another during the negotiation, either by hiding some of their tasks or by ...

Various researchers in Distributed Artificial Intelligence (DAI) have suggested that it would be worthwhile to isolate "aspects of cooperative behavior," general rules that would cause agents to act in ways conducive to cooperation. The hypothesis is that when agents act in certain ways (e.g., share information, act in predictable ways, defer globally constraining choices), it will be easier for them to carry out effective joint action. Another kind of cooperative behavior, less explored in the literature, is when agents independently alter the environment to make it easier for everyone to function effectively. Cooperative behavior of this kind might be to put away a hammer that one finds lying on the floor, knowing that another agent will be able to find it more easily later on. In this paper we are concerned with cooperation of this latter type, state-changing behavior that improves the environment for everyone. We examine the effect that a specific "cooperativity rule" has on agents...

Negotiation among multiple agents remains an important topic of research in Distributed Artificial Intelligence (DAI). Most previous work on this subject, however, has focused on bilateral negotiation, deals that are reached between two agents. There has also been research on n-agent agreement which has considered "consensus mechanisms" (such as voting), that allow the full group to coordinate itself. These group decision-making techniques, however, assume that the entire group will (or has to) coordinate its actions. Sub-groups cannot make sub-agreements that exclude other members of the group. In some domains, however, it may be possible for beneficial agreements to be reached among sub-groups of agents, who might be individually motivated to work together to the exclusion of others outside the group. This paper considers this more general case of n-agent coalition formation. We present a simple coalition formation mechanism that uses cryptographic techniques for subadditive Task Or...

We present a general theory that captures the relationship between certain domains and negotiation mechanisms. The analysis makes it possible to categorize precisely the kinds of domains in which agents find themselves, and to use the category to choose appropriate negotiation mechanisms. The theory presented here both generalizes previous results, and allows agent designers to characterize new domains accurately. The analysis thus serves as a critical step in using the theory of negotiation in realworld applications. We show that in certain Task Oriented Domains, there exist distributed consensus mechanisms with simple and stable strategies that lead to efficient outcomes, even when agents have incomplete information about their environment. We also present additional novel results, in particular that in concave domains using all-or-nothing deals, no lying by an agent can be beneficial, and that in subadditive domains, there often exist beneficial decoy lies that do not require full i...

This paper lays part of the groundwork for a domain theory of negotiation, that is, a way of classifying interactions so that it is clear, given a domain, which negotiation mechanisms and strategies are appropriate. We define State Oriented Domains, a general category of interaction. Necessary and sufficient conditions for cooperation are outlined. We use the notion of worth in an altered definition of utility, thus enabling agreements in a wider class of joint-goal reachable situations. An approach is offered for conflict resolution, and it is shown that even in a conflict situation, partial cooperative steps can be taken by interacting agents (that is, agents in fundamental conflict might still agree to cooperate up to a certain point). A Unified Negotiation Protocol (UNP) is developed that can be used in all types of encounters. It is shown that in certain borderline cooperative situations, a partial cooperative agreement (i.e., one that does not achieve all agents' goals) might be ...