Transport is one of the most important economic activities in our society. Economic growth has led to an increase in transport demand in both developed and emerging countries. This activity has grown in importance and complexity due to increased population and land area of this population. This increased demand in regions and transport modes, has led to the demand exceeds transport services supply, causing to reappear old problems such as congestion, pollution, accidents, financial deficits, etc. Traditionally, the main effort of the transportation planners has been the building of new transport infrastructures to improve its supply.
The design process of new network highways is carried out in a sequence of successive stages, in which the results of each stage are the inputs (data) of the next stage. In the early stages, the selection of a new highway corridor is a multicriteria decision-making process which should take into account social aspects, the spatial demand distribution, environmental and economic factors evaluated and weighted by a large number of alternatives of the highway corridor. In later stages, the preliminary locations are refined, taking into account other relevant aspects of the problem.
In this thesis, the highway network design problem is dealt with in the context of strategic planning.
Three methodologies based on optimization models for locating highway network have been developed, taking into account in all cases the construction costs and the demand pattern via the so-called origin-destination demand matrix in the study area. In particular:
- Chapter 3 develops a continuous location methodology to determine a set of potential highway corridors, so as to maximize the demand capture and subject to budgetary constraints.
- A continuous bi-level methodology to expand an existing highway network is developed in Chapter 4. In this model the user's behaviour is taken into account in the design process.
-Discrete location models for expansion of an existing highway network are stated in Chapter 5. The problem of improving an existing road network is addressed by an extension of fixed cost, capacitated and multicommodity network design model.
These models use geographic information to estimate the construction costs (the right-of-way cost, geological characteristics of the terrain, earthwork, pavement, etc). Some elements of these models are defined as ``black boxes'' that receive input from databases and process it by computational frameworks, that causing a large algorithmic complexity in the proposed models, require the use of heuristic algorithms for the optimization. In this thesis we have used: Particle Swarm Optimization, Simulated Annealing, Modified Simulated Annealing, based on the Cholesky decomposition, Simplex Nelder-Mead method and Scatter Search.
Computational experiments for the real case study of Castilla-La Mancha have been developed for each of the three proposed methodologies.