Date of this Version
R. A. Rajabzadeh, Measuring Autonomy And Solving General Stabilization Problems With Multi-Agent Systems, Lincoln: University of Nebraska, Lincoln, 2014.
Many distributed complex problems address a particular form of resource scheduling where proper resource management can cut costs by stabilizing a set of stochastic fluctuating parameters. Wireless sensor network communication, supply chain management, stock trading, intelligent traffic management, and smart grid systems are examples of these problems. Among the various solutions, a common strategy often used to address this type of problems is fluctuation reduction via resource buffering combined with load shifting. Respectively, stable wireless communication, demand for supplies, liquidity, traffic speed, and power demand reduce cost and can be achieved by properly managing sensor data buffers, warehouses, capital, distance between vehicles, and power storage units. Although on the surface, the differences between such problems appear to warrant completely different multi-agent solutions, they can be rephrased or approximated in common terms that enable the generalization of various solutions and techniques.
This thesis is concerned with generalizing fluctuation reduction problems and their solutions. To that end, this thesis defines the fluctuation problem class in a multiagent framework, provides a general solution, applies the general solution to the smart grid problem, investigates the solution dynamics with respect to common multi-agent system techniques, and finally defines a set of solution approach autonomy measurements. The resulting conceptual framework and applied investigation are directed at synthesizing currently disparate MAS research efforts which address fluctuation stabilization.
Advisor: Leen-Kiat Soh