Cost Effective and Survivable Cabling Design under Major Disasters

大災難下具有成本效益和生存能力的佈線設計

Student thesis: Doctoral Thesis

View graph of relations

Author(s)

  • Cong CAO

Related Research Unit(s)

Detail(s)

Awarding Institution
Supervisors/Advisors
Award date2 Dec 2015

Abstract

Survivability is one of the most important requirements in the design of telecommunication networks, as we are increasingly relying on the Internet and service interruptions are very costly. Unfortunately, unpredictable disasters, such as earthquakes and nuclear explosions, pose significant large-scale threats to the cables of telecommunication networks. In addition, other modern infrastructures, such as fuel or gas pipelines, electric power lines, roads and railway tracks, can also be adversely affected by various of disasters with grave and costly consequences. Therefore, achieving a low probability of “cable” or “network” damage is important for the proper operations of modern infrastructure.
This thesis addresses the important problem of how to optimize cabling between a given set of nodes positioned in a 2-dimensional plane under geographically correlated failures. We aim to either minimize cost subject to maintaining a required network survival probability, or alternatively, maximize the network survival probability subject to a given cost budget, in an event of a disaster. Our approach is to develop tractable disaster models based on the probability distributions of the disaster location and the disaster effects on the cables. Then, we consider various sets of network topology (or single cable shape) alternatives. For each set of alternatives, we either derive explicit expressions for the network survival probability and the cost based on the disaster models, or rely on simulations to evaluate them. For a given set of nodes, we aim to provide a cost effective and survivable cabling design solution.
As a first step, we start with the fundamental problem of how to connect two nodes with a single cable that crosses an earthquake fault line. We formulate a multi-objective optimization problem, with cost and cable break probability as the two objectives. We consider two important and meaningful alternatives of cable shape and provide intuitive justifications for choosing them. For each alternative of cable shape, we determine the Pareto front for the two objectives.
Next, we consider the problem of how to lay multiple cables between a pair of nodes in order to maintain network survivability under major disasters. We formulate a multi-objective optimization problem, with total cable cost and network survival probability as the two objectives. Three alternatives of network topologies are considered in order to reduce the probability that all the cables are simultaneously damaged in a single disaster. For each alternative network topology, we provide the Pareto front for the two objectives. We also discuss various scenarios for different disasters and cable break probability functions.
Finally, we extend the discussion to a network with multiple nodes. We first design a survivable topology for N nodes where the N nodes can form a convex polygon. Then we present a heuristic algorithm for designing cable topologies of a general N-node network and illustrate the algorithm with two numerical examples. This thesis provides a methodology for discussing and solving a real world problem of how to lay cables.