Multiobjective Optimization Models and solution Methods for Planning land Development using Minimum Spanning Trees, Lagrangian Relaxation and Decomposition Techniques

Abstract

Title: MULTIOBJECTIVE OPTIMIZATION MODELS AND SOLUTION METHODS FOR PLANNING LAND DEVELOPMENT USING MINIMUM SPANNING TREES, LAGRANGIAN RELAXATION AND DECOMPOSITION TECHNIQUES José Alberto Faria, Doctor of Philosophy, 2005 Directed By: Professor Steven A. Gabriel Department of Civil and Environmental Engineering and Applied Mathematics and Scientific Computation Program University of Maryland The land development problem is presented as the optimization of a weighted average of the objectives of three or more stakeholders, subject to develop within bounds residential, industrial and commercial areas that meet governmental goals. The work is broken into three main sections. First, a mixed integer formulation of the problem is presented along with an algorithm based on decomposition techniques that numerically has proven to outperform other solution methods. Second, a quadratic mixed integer programming formulation is presented including a compactness measure as applied to land development. Finally, to prevent the proliferation of sprawl a new measure of compactness that involves the use of the minimum spanning tree is embedded into a mixed integer programming formulation. Despite the exponential number of variables and constraints required to define the minimum spanning tree, this problem was solved using a hybrid algorithm developed in this research. MULTIOBJECTIVE OPTIMIZATION MODELS AND SOLUTION METHODS FOR PLANNING LAND DEVELOPMENT USING MINIMUM SPANNING TREES, LAGRANGIAN RELAXATION AND DECOMPOSITION TECHNIQUES

158 Figures and Tables

Cite this paper

@inproceedings{Faria2005MultiobjectiveOM, title={Multiobjective Optimization Models and solution Methods for Planning land Development using Minimum Spanning Trees, Lagrangian Relaxation and Decomposition Techniques}, author={Jose Alberto Faria}, year={2005} }