Dissertation > Excellent graduate degree dissertation topics show

Multi-period Network Design Problem under Demand Uncertainties

Author: HuWei
Tutor: XiuNaiHua
School: Beijing Jiaotong University
Course: Operational Research and Cybernetics
Keywords: Stochastic Demand Random network Network design Multi-stage
CLC: F51
Type: Master's thesis
Year: 2010
Downloads: 92
Quote: 0
Read: Download Dissertation

Abstract

Transport infrastructure investment is critical, strategic decision-making in urban planning and economic development, occupies a very important one seat. Most traffic network project with the same three characteristics: high costs, partially or completely reversible and investment stages of various uncertainties. Most of the current transportation investment decision-making research for the future state of the single-point estimates. This decision did not take into account the uncertainty of the future, and thus when the future state is not expected, there is no way to take measures to improve the system, this will bring huge economic losses to society. In this paper, we consider a network design problem under the random demand. Uncertainty of traffic demand has led to the uncertainty of the network parameters of the entire transport network, namely the deterministic transport network become random traffic network. Random traffic network standard traffic equilibrium condition is no longer applicable, because the link cost is uncertain. And, when certain sections traffic flow to meet the equilibrium condition, the desired section of transport costs does not necessarily satisfy the equilibrium conditions. Elastic demand situation, the the uncertain transportation costs will in turn affect the traffic demand, thus affecting the entire transportation network, and thus the equilibrium conditions of the standard definition is no longer applicable. Nevertheless, recently proposed in the literature random elastic demand transportation network equilibrium conditions, but the lack of the dominant random traffic demand and random traffic costs the exact relationship between random demand modeling, so did not give the exact method of calculation. By modifying the traditional MSA algorithms to overcome the lack of the dominant demand function relationship difficulties, random elastic demand balance calculation method of road network users. Finally, this paper presents a numerical algorithm in a classic network cases, numerical results demonstrate the validity of the theory and algorithms. On this basis, this paper studies the multi-stage random network design problems of the transport network, mainly considering road pricing scheme. This paper takes into account not only the the random traffic network literature in the recent widespread use of short-term traffic demand uncertainty, and the introduction of the phase of the long-term growth in traffic demand uncertainty. In this paper for the long-term uncertainty, a random scenario tree method (Stochastic Scenario-tree), extends the traditional scenario tree. Further introduction of flexible dependencies between the stage, to some extent reflects the impact of land use, and more realistic. Based on the above, this paper established a network design model of multi-stage random traffic network, and existence theorems are given to understand, as well as the genetic algorithm-based computing framework and gives numerical examples.

Multi-period Network Design Problem under Demand Uncertainties

Abstract

Related Dissertations