Combining activity analysis models

Abstract

This thesis presents two methods for combining activity analysis models interacting together. In order to obtain an equilibrium solution which is equivalent to the optimum solution of the entire system, mathematical programming decomposition/integration methodologies are investigated. The firstly developed solution scheme is named hybrid approach, since it combines two conventional decomposition methods, say, price directive and resource directive approaches. The convergence of the hybrid decomposition algorithm is proved. The second part of the thesis proposes a decomposition scheme without any master problems. A new formulation and solution procedures are presented. The convergence properties are considered relative to the hybrid method with a master problem.