Jaroslav Barta

A knowledge-based approach to coalition formation

Establishing coalitions among humanitarian aid organizations is a challenge. The CPlanT (Coalition Planning Tool) system uses a multi-agent knowledge-based approach to reduce complexity and communication traffic, while also protecting stakeholder privacy.

Role of acquaintance models in agent’s private and semi-private knowledge disclosure

Abstract The organizational architecture of the multi-agent systems and the structure of social knowledge that the members of the community administer are critical factors for assuring such patterns of information exchange that keep agents private knowledge confidential. In this paper we will introduce the concept of agents’ private and semi-private knowledge and we will explain the difference between the alliance – a semi-permanent organizational structure and a coalition – a goal-oriented, non-permanent organizational structure. We will provide the reader with an analysis on how does the agents’ social knowledge, stored in the tri - base acquaintance model , contributes to permanent confidentiality of agents’ private knowledge, preferences, decision making models, resources, etc. The study has been experimentally verified in the domain of planning for humanitarian relief operations within a high number of hardly collaborating and vaguely linked non-governmental organizations is a challenging problem.

Meta-reasoning for Agents’ Private Knowledge Detection

Agent’s meta-reasoning is a computational process that implements agent’s capability to reason on a higher level about another agent or a community of agents. There is a potential for meta-reasoning in multi-agent systems. Meta-reasoning can be used for reconstructing agents’ private knowledge, their mental states and for prediction of their future courses of action. Meta-agents should have the capability to reason about incomplete or imprecise information. Unlike the ordinary agents, the meta-agent may contemplate about the community of agents as a whole. This contribution presents application of the meta-reasoning process for the agent’s private knowledge detection within the multi-agent system for planning of humanitarian relief operations.

CPlanT: Coalition Planning Multi-Agent System for Humanitarian Relief Operations

Planning humanitarian relief operations within a community of a high number of hardly collaborating and vaguely linked non-governmental organizations is a challenging problem. In this article we present an agent-oriented approach to modeling and supporting distributed decision making of agents - the humanitarian relief providers. Owing to the very special nature of this specific domain, where the agents may eventually agree to collaborate but are very often reluctant to share their knowledge and resources, we have combined classical negotiation mechanisms with the acquaintance models and social knowledge techniques.

Distributed Branch and Bound Algorithm in Coalition Planning

We suggest the Distributed Branch and Bound Algorithm intended for solving problems with exponential complexity e.g. coalition planning. This technique consists of three parts: the state space distribution, effective search algorithm and partial solution expansion. We will describe CPlanT - an OOTW (operations other than war) coalition planning multi-agent system where this approach has been exploited.