Multiagent Metareasoning through Organizational Design

In this paper we outline a general approach to the study of metareasoning, not in the sense of explicating the semantics of explicitly speci ed meta-level control policies, but in the sense of providing a basis for selecting and justifying computational actions. This research contributes to a developing attack on the problem of resource-bounded rationality, by providing a means for analysing and generating optimal computational strategies. Because reasoning about a computation without doing it necessarily involves uncertainty as to its outcome, probability and decision theory will be our main tools. We develop a general formula for the utility of computations, this utility being derived directly from the ability of computations to a ect an agent's external actions. We address some philosophical di culties that arise in specifying this formula, given our assumption of limited rationality. We also describe a methodology for applying the theory to particular problem-solving systems, and provide a brief sketch of the resulting algorithms and their performance. two reviewers for their valuable comments and suggestions.

2009

In this paper we present an approach to design an Organized Multiagent Systems (OMAS) for teamwork. We use a general formal model for OMAS that employs the notion of organizational mechanisms. The purpose of such mechanisms is influencing the behaviour of the agents towards more effectiveness with regard to some objectives. To achieve our goal we use Markov Decision Processes (MDPs) as a framework to design the organizational mechanisms.

Recent research has shown how an organization can influence a decision-theoretic agent by replacing one or more of its model components (transition/reward functions, action/state spaces, etc.), and how each of these influences impacts the agent's decision-making performance. This paper delves more precisely into exactly which parts of an agent's model should be organizationally influenced, and asserts a broader principle for delineating what aspects of an agent's behavior an organization should be sanctioned to influence. We present a formal framework for specifying factored organizational influences and incorporating them into agents' decision models, and empirically demonstrate that organizational specifications based on our proposed principle outperform the alternatives. We further describe an algorithm for automating the organizational-design process that is inspired by this principle, and demonstrate empirically that its organizational designs are both intuitive...

Southeastern Symposium on System Theory, 2001

Agent based computing offers the ability to decentralize computing solutions by incorporating autonomy and intelligence into cooperating, distributed applications. It provides an effective medium for expressing solutions to problems that involve interaction with real-world environments and allows modelling of the world state and its dynamics. This model can be then used to determine how candidate actions affect the world and

Lecture Notes in Computer Science, 2005

An autonomous agent may largely benefit from its ability to reconstruct another agent's reasoning principles from records of past events and general knowledge about the world. In our approach, the meta-agent maintains a first-order logic theory, called the community model, yielding predictions about other agents' decisions. In this contribution we introduce a query-based collective reasoning process where the semi-collaborative meta-agents use active learning technique to improve their models. We provide empirical results that demonstrate the viability of the concept and show the benefits of collective meta-reasoning.

2004

Several agent-oriented methodologies have been proposed over the last few years. Unlike the object-oriented domain and unfortunately for designers, most of the time, each methodology has its own purposes and few standardization works have been done yet, limiting the impact of agent design on the industrial world. This paper tries to find a means to unify three existing methodologies -ADELFE, Gaia and PASSI -by studying their meta-models and the concepts related to them. Comparing a certain number of features at the agent or system level (such as the agent structure, its society or organisation, its interactions capacities or how agents may be implemented) has enabled us to draw up a first version of a unified meta-model proposed as a first step towards interoperability between agent-oriented methodologies.

2004

Effects-based design of robust organizations seeks to synthesize an organizational structure and its strategies (resource allocation, task scheduling and coordination), to achieve the desired effect(s) in a dynamically changing mission environment. In this paper, we model the dynamic system associated with the mission environment (e.g., environment faced by a joint task force with a military objective

Proceedings of the 5th International Conference on Agents and Artificial Intelligence, 2013

Self-organizing multi-agent systems provide a suitable paradigm for agents to manage themselves. We demonstrate a robust, decentralized approach for structural adaptation in explicitly modelled problem solving agent organizations. Based on self-organization principles, our method enables the agents to modify their structural relations to achieve a better completion rate of tasks in the environment. Reasoning on adaptation is based only on the agent's history of interactions. Agents use the history of tasks assigned to their neighbours and completion rate as a measure of evaluation. This evaluation suggests the most suitable agents for reorganization (Meta-Reasoning). Our Selective-Adaptation has four different approaches of Meta-Reasoning, which are 1) Fixed Approach, 2) Need-Based Approach, 3) Performance-Based Approach, and 4) Satisfaction-based Approach along with a Reorganization approach, which needs less data but makes better decisions.

2005

Despite all the research done in the last years on the development of methodologies for designing MAS, there is no methodology suitable for the specification and design of MAS in complex domains where both the agent view and the organizational view can be modeled. Current multiagent approaches either take a centralist, static approach to organizational design or take an emergent view in which agent interactions are not pre-determined, thus making it impossible to make any predictions on the behavior of the whole systems.

The Journal of Logic Programming, 2000

There are numerous applications where an agent a needs to reason about the beliefs of another agent, as well as about the actions that other agents may take. In [T. Eiter, V.S. Subrahmanian, G. Pick, Heterogeneous Active Agents, I: Semantics, Arti®cial Intelligence 108 (1±2) (1999) 179±255] the concept of an agent program is introduced, and a language within which the operating principles of an agent can be declaratively encoded on top of imperative data structures is de®ned. In this paper we ®rst introduce certain belief data structures that an agent needs to maintain. Then we introduce the concept of a Meta Agent Program (map), that extends the framework of Refs. [T. Eiter, V.S. Subrahmanian, Heterogeneous Active Agents, II: Algorithms and Complexity, Arti®cial Intelligence 108 (1±2) (1999) 257±307; loc. cit.] so as to allow agents to perform metareasoning. We build a formal semantics for maps, and show how this semantics supports not just beliefs agent a may have about agent b 's state, but also beliefs about agents b 's beliefs about agent c 's actions, beliefs about b 's beliefs about agent c 's state, and so on. Finally, we provide a transansation that takes any map as input and converts it into an agent program such that there is a one±one correspondence between the semantics of the map and the semantics of the resulting agent program. This correspondence allows an implementation of maps to be built on top of an implementation of agent programs.