An approach to the highway alignment development process using genetic algorithm based optimisation

Ahmad Al-Hadad, Botan (2011) An approach to the highway alignment development process using genetic algorithm based optimisation. PhD thesis, University of Nottingham.

[img]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (15MB) | Preview

Abstract

Highway alignment development is recognised as a non-linear constrained optimisation problem. It is affected by many economical, social, and environmental factors subject to many design constraints. The highway alignment development process is therefore considered complex but very important.

Highway alignment development is about finding an optimum alignment solution between two termini points in a 3D space, subject to several constraints. The development process using the current method is considered complex because of the number of the design elements involved, their interactions, and the formulations required to relate them to a realistic highway alignment. Moreover, it is considered that an alignment, generated using the existing method, results in a sub-optimal solution. This is due to the fact that the two alignments (horizontal and vertical alignments) are found in two independent stages and from only a handful number of alternative evaluations.

This research reports on a new approach for improving the process of highway alignment development by utilising modern technologies. It proposes a novel design approach, as an alternative to the existing method, for highway alignment development in a three-dimensional space (considering the horizontal and vertical alignments simultaneously). It describes a method for highway alignment development through station points. Station points, as points along the centre line of alignment which are defined by their X, Y, and Z coordinates, are used to define the alignment configuration. The research also considers the implications of access provision (in term of junctions) and their locations on highway alignment. The environmental factors (noise and air pollution in terms of proximity distance) and accessibility (user and link construction costs in terms of access costs) are embedded in the formulations required to represent junctions in the model.

The proposed approach was tested through the development of a genetic algorithms based optimisation model. To achieve this, several algorithms were developed to perform the search. The evaluation of the solutions was handled by a fitness function that includes construction (length), location (land acquisition, environmentally sensitive areas, and soil condition), and earthwork (fill and cut material) dependent costs. Other forms of costs that are quantifiable can also be incorporated within the fitness function. The critical constraints, believed important for realistic alignments (horizontal curvature, vertical curvature, and maximum gradient) are also dealt with within the model formulation.

The experimental results show that the problem of highway alignment can be better represented using the concept of station points, by which better alignment solutions (global or near global solutions) were achieved. It was also shown that the alignment development process could be simplified through the use of station points, resulting in the efficient evaluation of more alternatives. Furthermore, the results conclude that a highway alignment cannot be optimum unless it is simultaneously optimised with junctions. Further investigations and development are also recommended for future studies.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Mawdesley, M.J.
Stace, L.R.
Keywords: Roads, design and construction, genetic algorithms
Subjects: T Technology > TE Highway engineering. Roads and pavements
Faculties/Schools: UK Campuses > Faculty of Engineering > Department of Civil Engineering
Item ID: 12259
Depositing User: EP, Services
Date Deposited: 09 Mar 2012 10:26
Last Modified: 06 May 2020 11:31
URI: https://eprints.nottingham.ac.uk/id/eprint/12259

Actions (Archive Staff Only)

Edit View Edit View