| Information technology is the key driver for many dramatic changes in the structure of organizations and new types of electronic markets. This dissertation provides a new framework for market-based information systems design. It consists of three essays.; The first essay "A Market-Based Optimization Algorithm for Distributed Systems" studies the allocative roles of markets in a distributed environment where different agents have different information, act in their own interests, and compete for scarce resources such as CPU time, storage space, or bandwidth. We develop a market-oriented decomposition algorithm to effectively solve a competitive resource allocation problem in a double auction market environment. The global system optimal solution can be found via iterative trading of bundled resources synchronously or asynchronously in the marketplace in a finite number of steps. We provide a market-based paradigm for decentralized problem-solving that can be easily implemented to support real-time optimization of distributed systems.; The second essay focuses on the implementation issues of our new algorithm. In the essay "Implementing a Market-Based Optimization Algorithm for Distributed Systems", we sought to understand how various exogenous factors and strategic agent behavior might affect the market's convergence speed and solution quality. In particular, we test how different problem structures, different initial endowments, and different strategic interaction among agents will have an effect on the system performance, and the corresponding social welfare impacts. We derive some important insights in implementing our algorithm in real auction market environment, thus providing practical managerial implications.; In the third essay, "Supply Chain Information Sharing in Macro Prediction Markets", we analyze the economic rationale for introducing new markets in the supply chain information systems. A critical element in our model is the construction of a macro prediction market, in which dispersed information from a variety of sources are effectively collected, efficiently processed and widely disseminated. Our analysis suggests that properly designed markets can be used to align different organization's incentives to share their private information in a mutually beneficial way so that coordination efficiency can be greatly improved. |
