UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Genetic algorithms for multi-objective optimization in water quality management under uncertainty Tolson, Bryan Antony

Abstract

This thesis demonstrates the combined usage of a number of novel approaches and techniques in the multi-objective management of water quality systems under uncertainty. The First-Order Reliability Method (FORM) is used to estimate the risk-based system performance indicators of reliability, vulnerability, and resilience that provide measures of the frequency, magnitude and duration of the failure of water resource systems, respectively. FORM accuracy and efficiency for performance indicator estimation is compared extensively with Monte Carlo Simulation (MCS). Genetic Algorithms (GAs) are demonstrated as a robust optimization technique by solving various multi-objective water quality management models that optimize the performance indicators and the total point source waste treatment cost. In addition, the Tradeoff Surface Representation (TSR) Algorithm is incorporated as a general multi-objective technique for accurate and efficient identification of convex tradeoff surfaces. The Willamette River Basin in Oregon, USA is utilized as the water quality management case study for the demonstration of all techniques. The performance indicators are estimated with respect to meeting dissolved oxygen (DO) standards and ambient DO is simulated using a QUAL2E water quality response model. Results show that FORM estimates of the performance indicators, while significantly less accurate than MCS estimates, seem to provide reasonable results when utilized within the multiobjective water quality management models. A comparison of FORM and MCS shows that while FORM is more efficient relative to MCS, the difference in efficiencies is significantly less than previously reported in the literature. The TSR Algorithm, in comparison with the commonly used Constraint Method for multi-objective tradeoff curve generation, is shown to produce a superior representation of the tradeoff curve. Furthermore, the TSR Algorithm is also shown to produce a maximum amount of auxiliary information regarding the bounds on the location of the tradeoff curve between tradeoff points.

Item Media

Item Citations and Data

Rights

For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use.