A Market-Based Optimization Algorithm for Distributed Systems
提出一种市场分解方法，通过主问题解决市场匹配、子问题解决代理捆绑确定，在有限交易内实现系统最优，适用于分布式系统的实时优化。
In this paper, a market-based decomposition method for decomposable linear systems is developed. The solution process iterates between a master problem that solves the market-matching problem, and subproblems that solve the agents’ bundle-determination problems. Starting from any initial price and feasible allocation, system optimality can be achieved under a dynamic market-trading algorithm in a finite number of trades. The final market-clearing prices are discovered by this market trading and an efficient allocation is achieved by direct, wealth-improving resource exchanges among self-interested agents. Certain types of strategic behavior by the agents and a dealer in the marketplace are studied as well. Our proposed market mechanism addresses price dynamics, incentive issues, and economic transactions of real-world, distributed decision-making situations more realistically than traditional decomposition approaches. In addition, it can be operated in both synchronous and asynchronous environments. We provide a market-based paradigm for decentralized problem solving and information processing that can be easily implemented to support real-time optimization of distributed systems.