ABSTRACT DATA TYPES FOR KNOWLEDGE REPRESENTATION AND SPECIFICATION OF MULTI-AGENT SYSTEMS
Ключевые слова:
multiagent systems and algorithms, assignment problem, safety and progress properties, algorithm verification, privacy, anonymity, Social Software.Аннотация
Distributed system is a group of decentralized interacting executers. Distributed algorithm is the
communication protocol for a distributed system that transforms the group into a team to solve some task. Multiagent
system is a distributed system that consists of autonomous reactive agents, i.e. executers which internal states can be
characterized in terms Believes (B), Desires (D), and Intentions (I). Multiagent algorithm is a distributed algorithm
for a multiagent system.
The article discusses the basic concepts of agents and multi-agent systems. Also, two problems of multi-agent
algorithms for representing knowledge in the context of Social Software Engineering are considered. A number of
new multi-agent algorithms are presented, and their correctness is proved. The main characteristics of agents are
provided, such as autonomy, proactivity, social ability, and reactivity; also, agents can have such additional
characteristics as persistence, reasonability, performance, mobility, personality, and rationality. A number of new
multi-agent algorithms are presented, and their correctness is proved. Two statements have been proved for solving
RAM and MRP problems. This time we address a social issue of agent anonymity and privacy in these algo-rithms.
