Development of an interval quadratic programming water quality management model and its solution algorithms

Abstract Interval quadratic programming (IQP) is one of the most popular programming techniques for addressing the uncertainties associated with nonlinear environmental management problems. However, efficient algorithms for solving IQP problems in water quality management (WQM) have not been well studied. In this study, an IQP model is developed for WQM under uncertainty. Three solution algorithms, including a piecewise linear approximation (PLA) method, a derivative algorithm (DEA) and a duality-based algorithm (DUA) are proposed for solving the IQP-WQM problem. The developed model and the corresponding solution algorithms are applied to a hypothetic WQM problem to demonstrate their applicability. The results show the lower bounds of the total cost obtained by three algorithms have a relationship of f D U A − = f D E A − ≤ f P L A − , while that of the upper bounds is f D E A + ≤ f P L A + ≤ f D U A + . Moreover, the sensitivity analysis shows that no matter how the IQP-WQM model is solved, the model response is consistent among the three solution algorithms. The results indicate that all of the three algorithms can efficiently deal with quadratic programming problems under uncertainties in the format of intervals. Comparison among the three algorithms shows that DUA provides interval solutions with wider ranges than the other two methods, and it requires less computational efforts than DEA. It is also found that PLA is more flexible and might require lower computational efforts for large scale problems. This study could provide useful decision support for effective WQM.