作者: Yi-Chen E. Yang , Ximing Cai , Dušan M. Stipanović
DOI: 10.1029/2008WR007634
关键词: Distributed computing 、 Multi-agent system 、 Watershed 、 Optimization methods 、 Decision process 、 Scheme (programming language) 、 Watershed management 、 Decentralized optimization 、 Factor (programming language) 、 Engineering 、 Management science
摘要: [1] A watershed can be simulated as a multiagent system (MAS) composed of spatially distributed land and water users (agents) within common defined environment. The is characterized by decision processes at the agent level with coordination mechanism organizing interactions among individual level. This paper presents decentralized (distributed) optimization method known constraint-based reasoning, which allows agents in an MAS to optimize their behaviors over various alternatives. incorporates all agents' objectives through interaction scheme, ith optimizes its objective selected priority for collaboration forwards solution consequences that interact it. Agents are allowed determine how important own comparison constraints, using local interest factor (βi). A large βi value indicates selfish who puts high on benefit ignores requirements. bottom-up problem-solving approach mimics real-world management problems better than conventional “top-down” methods it assumed will completely comply any recommendations coordinator makes. applied steady state hypothetical three off-stream human agents, one in-stream (reservoir), two ecological agents.