A dynamic model of integrated production-transportation operations

The Integrative Freight Market Simulation (IFMS) aims at developing a comprehensive urban freight transportation demand model, which includes three key dimensions of the freight system: (1) financial, by focusing on profit maximization; (2) commodity flows, that represent demand; and (3) vehicle trips, the logistical response of the supply side to satisfy consumer demands. This thesis focuses on the development of analytical formulation to estimate urban freight-related flows for integrated shipper-carrier operations in the IFMS framework. It was assumed that the market is composed of two key agents: suppliers and consumers. The suppliers perform both shipping and transportation functions. The suppliers are profit maximizers who produce, sell and deliver a generic commodity. The consumers are receivers with a fixed demand for the same commodity. The problem studies how the suppliers adjust production level, profit margin, and routing strategy in response to the competition, until Walrasian equilibrium is reached satisfying the following conditions: (1) the suppliers' profits are maximized, they are not able to earn more by changing their own business strategies; (2) the consumer demands are satisfied; and (3) the market clears, and no inventory is left.;The solution methodology used heuristics to iteratively solve the problem to ensure equilibrium. At each time period, the problem consisted of three stages. The first stage, termed "target", considered the sequence of decisions by which the suppliers decide how much to provide and deliver at what price based on their expectations of how much to sell in the market, this included the decisions of production level, profit margin, preliminary routing, as well as the estimation of total cost and price. A set of dynamics was used to help the suppliers make decisions on the productions for the next time period. The second stage was the market competition, which determines the amount that each consumer will purchase from the suppliers. This was formulated by a logit function of price to determine the market share. The third stage was termed "observed", during which the suppliers route the vehicles, make the deliveries, and estimate the performance metrics.;The algorithmic solutions were obtained for three types of dynamics of the suppliers' decisions on productions. For small problems, the suppliers tended to find their optimal routing and profit margins relatively quickly, thus the solutions generally converged to equilibrium relatively fast. At equilibrium, the suppliers with proximity to the market generally had lower marginal cost and charged higher profit margin, thus they yielded higher net profits; while those farther away usually gained lower profits and even lost money. Furthermore, if the consumers are less price sensitive, the suppliers could charge higher profit margins. For larger problems, the suppliers generally kept changing routing pattern and profit margin, and then it was more difficult to converge to equilibrium, in fact equilibrium may not exist. The analysis of market entry and exit indicates that, market proximity is important if there is significant locational difference. If locational difference is only marginal, then routing plays a significant role.