Article

Optimization of a Basin Network Using Hybridized Global Search Algorithms

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Over the last few decades, groundwater resources in many regions have been depleted at a higher rate than the underlying aquifers have been replenished. This imbalance has led water management agencies to consider managed aquifer recharge networks, in which infiltration basins are used to replenish the aquifers using previously uncaptured stormwater runoff. In this work, optimization methods were used to select parameter values to minimize the cost associated with constructing such a network while ensuring the network has the ability to supplement demands placed on the aquifer. The objective function considered incorporates land and construction costs, along with rewards for effective aquifer recharge, and constraints were incorporated to enforce capture of a minimum volume of stormwater runoff. Two hybridized global search algorithms were considered, one based on particle swarm optimization and the other on a genetic algorithm approach. Both methods returned solutions that were close in terms of minimal cost but varied in terms of individual basin sizes. Thus, the algorithms are able to aid decision makers by providing several cost-competitive solutions that can then be used to support a community dialogue.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... The third hybrid PSOPS algorithm was also applied to a basin network optimization problem (Beauregard et al. 2018). The results were compared to the results of PSO and GA that were published in Beauregard et al. (2018). ...
... The third hybrid PSOPS algorithm was also applied to a basin network optimization problem (Beauregard et al. 2018). The results were compared to the results of PSO and GA that were published in Beauregard et al. (2018). The PSOPS algorithm showed higher performance and more robustness compared with PSO with filter method and Genetic Algorithm (GA) with implicit filtering while also potentially finding the global minimum of the objective function in most trial runs. ...
... A recent paper compared the application of the hybrid PSO with filter method (Almomani 2012) and the hybrid GA with implicit filtering (Ritz 2017) to the optimization of basin networks (Beauregard et al. 2018). In Beauregard et al. (2018), the authors found that both the hybrid approach for PSO and GA were able to find configurations of Fig. 3 The performance of median function evaluations against from 1 to 10 Content courtesy of Springer Nature, terms of use apply. ...
Article
Full-text available
Particle swarm optimization (PSO) is one of the most commonly used stochastic optimization algorithms for many researchers and scientists of the last two decades, and the pattern search (PS) method is one of the most important local optimization algorithms. In this paper, we test three methods of hybridizing PSO and PS to improve the global minima and robustness. All methods let PSO run first followed by PS. The first method lets PSO use a large number of particles for a limited number of iterations. The second method lets PSO run normally until tolerance is reached. The third method lets PSO run normally until the average particle distance from the global best location is within a threshold. Numerical results using non-differentiable test functions reveal that all three methods improve the global minima and robustness versus PSO. The third hybrid method was also applied to a basin network optimization problem and outperformed PSO with filter method and genetic algorithm with implicit filtering.
Conference Paper
Full-text available
Over the last few decades, groundwater resources in many regions have been depleted at a faster rate than the underlying aquifers have been replenished. This imbalance has led water management agencies to consider managed aquifer recharge networks, where infiltration basins are used to replenish the aquifers using previously-uncaptured storm water runoff. In this work, we utilize optimization to evaluate the costs associated with constructing such a network and the ability of the network to meet demands placed on the aquifer. Our objective function incorporates land and construction costs, along with rewards for effective aquifer recharge. We enforce capture of a minimum volume of storm water runoff by penalizing the cost. We present results for two basin networks, one based on results from the literature and another based on a study of the Pajaro Valley region in California. The Pajaro Valley example is used as our realistic test case, and we use the analysis to suggest the viability of a managed aquifer recharge network in a particular sub-watershed associated with the area.
Article
Full-text available
Natural and anthropogenic conditions in the Gulf countries lead to substantial changes in groundwater (GW) dynamics. GW mound under a lagoon supplied by treated wastewater (TWW) from a municipal sewage treatment plant (STP) in Muscat, Oman, is studied using analytical and numerical modeling techniques. The water table mound in an unconfined aquifer, laterally bounded by the sea, receives recharge from a lagoon fed by a STP in a seasonally varying regime. The onset and duration of the filling period and, correspondingly, infiltration from the lagoon bed is dictated by the wastewater processing and TWW use. Following the period of recharge-induced growth, the GW mound decays when recharge ceases, and the void space above the water table becomes available for future injection in the next annual cycle of infiltration and TWW subsurface storage. The Green-Ampt model application illustrates that the first phase of the process (vertical infiltration with front propagating from the lagoon bed to the regional flat water table) is typically short (several hours). The second phase, when TWW gets in a direct hydraulic contact with GW, is the longest and with a mound spreading to hundreds of meters during winter months. No analytical solutions to this moving boundary-value problem are known and approximate models, based on the Boussinesq approximation and its linearizations, are utilized. The Hantush-type drawup of the water table in the vicinity of the lagoon is studied by the Watson analytical solution to the diffusion equation for a vertically averaged hydraulic head in an unbounded domain, with an instantaneous jump of the head over a disk-shaped zone. This drawup is compared with MODFLOW simulations, which take into account a finite size of the flow domain (catchment) and discharge into the sea through a constant-head segment. The Mishra-Guyonnet’s formula assumes a spatially uniform flux from the lagoon bed (disk in a model plain) into a growing mound and is used for calculation of the total discharge, which progressively decreases from the early stages of the Green-Ampt infiltration to a theoretically zero value if the second phase lasts indefinitely long. We assess characteristic times and phases of the mound dynamics during the rise and fall stages using regional hydrogeological parameters, site geometry, and typical patterns of wastewater disposal. Sensitivity analysis of the locus of the phreatic surface at a given instance and location is carried out, with storativity, hydraulic conductivity, undisturbed aquifer thickness and lagoon radius as input variables. For an example of an aquifer of saturated depth of 40 m, undisturbed vadose zone thickness of 10 m, saturated hydraulic conductivity of 10 m/day, effective porosity of 0.2, a circular lagoon size of 90 m and water depth there of 1 m we found that the infiltration front propagates to the water table in less than 0.2 days; after 40 days of a continuous mound growth in a hypothetical piezometer located 400 m from the lagoon the water table drawup was in the range of 2.0–2.7 m if the aquifer hydraulic parameters and lagoon size are varied 20 % from the average value. This approach will permit to plan more detailed site investigation, facilitate management decisions on the rate of water release to the lagoon, and foresee waterlogged areas.
Article
Full-text available
Retention basins are an effective and common means of flood control. Differing from detention, retention fully contains a particular storm event, using infiltration as the primary means of disposal. For a complex watershed with multiple subwatersheds; however, requiring basins to retain only flow from their contributing subwatershed may result in a suboptimal solution both in terms of cost and overall flood control. This paper introduces a design procedure for considering location and sizing of a network of retention basins within a watershed. It utilizes the discrete dynamic programming technique to determine the ideal locations and geometry for each basin. A design example demonstrates that this method may allow significant cost savings over traditional methods.
Article
Full-text available
Numerical tests indicate that Hortonian runoff mechanisms benefit from scaling effects that non-Hortonian runoff mechanisms do not share. This potentially makes Hortonian watersheds more amenable to physically based modelling provided that the physically based model employed properly accounts for rainfall distribution and initial soil moisture conditions, to which these types of model are highly sensitive. The distributed Hortonian runoff model CASC2D has been developed and tested for the US Army over the past decade. The purpose of the model is to provide the Army with superior predictions of runoff and stream-flow compared with the standard lumped parameter model HEC-1. The model is also to be used to help minimize negative effects on the landscape caused by US armed forces training activities. Development of the CASC2D model is complete and the model has been tested and applied at several locations. These applications indicate that the model can realistically reproduce hydrographs when properly applied. These applications also indicate that there may be many situations where the model is inadequate. Because of this, the Army is pursuing development of a new model, GSSHA, that will provide improved numerical stability and incorporate additional stream-flow-producing mechanisms and improved hydraulics. Copyright © 2002 John Wiley & Sons, Ltd.
Article
Full-text available
1] The term distributed model is widely applied to describe hydrologic models that can simulate broad classes of pathways of water through space, e.g., overland flow, saturated groundwater flow, and/or unsaturated flow in the vadose zone. Because existing distributed modeling approaches differ substantially from one another, we present a common framework from which to compare the many existing hillslope-and catchment-scale models. To provide a context for understanding the structure of the current generation of distributed models, we briefly review the history of hydrologic modeling. We define relevant modeling terms and describe common physical, analytical, and empirical approaches for representing hydrologic processes in the subsurface, surface, atmosphere, and biosphere. We then introduce criteria for classifying existing distributed models based on the nature of their process representation, solution scheme, coupling between the surface and subsurface, and treatment of space and time. On the basis of these criteria we describe 19 representative distributed models and discuss how process, scale, solution, and logistical considerations can be incorporated into model selection and application. Citation: Kampf, S. K., and S. J. Burges (2007), A framework for classifying and comparing distributed hillslope and catchment hydrologic models, Water Resour. Res., 43, W05423, doi:10.1029/2006WR005370.
Article
Full-text available
Many optimization problems are characterized by expensive objective and/or constraint function evaluations paired with a lack of derivative information. Direct search methods such as generating set search (GSS) are well understood and efficient for derivative-free optimization of unconstrained and linearly constrained problems. This paper presents a study of heuristic algorithms that address the more difficult problem of general nonlinear constraints where derivatives for objective or constraint functions are unavailable. We focus on penalty methods that use GSS to solve a sequence of linearly constrained problems, numerically comparing different penalty functions. A classical choice for penalizing constraint violations is ℓ 2 2 , the squared ℓ 2 norm, which has advantages for derivative-based optimization methods. In our numerical tests, however, we show that exact penalty functions based on the ℓ 1 ,ℓ 2 , and ℓ ∞ norms converge to good approximate solutions more quickly and thus are attractive alternatives. Unfortunately, exact penalty functions are nondifferentiable and consequently degrade the final solution accuracy, so we also consider smoothed variants. Smoothed-exact penalty functions are attractive because they retain the differentiability of the original problem. Numerically, they are a compromise between exact and ℓ 2 2 , i.e., they converge to a good solution somewhat quickly without sacrificing much solution accuracy. Moreover, the smoothing is parameterized and can potentially be adjusted to balance the two considerations. Since our focus is on optimization problems that are characterized by expensive function evaluations, reducing the number of function evaluations is paramount, and the numerical results of this paper suggest that exact and smoothed-exact penalty functions are well-suited to this task.
Article
Full-text available
Efficient and powerful methods are needed to overcome the inherent difficulties in the numerical solution of many simulation-based engineering design problems. Typically, expensive simulation codes are included as black-box function generators; therefore, gradient information that is required by mathematical optimization methods is entirely unavailable. Furthermore, the simulation code may contain iterative or heuristic methods, low-order approximations of tabular data, or other numerical methods which contribute noise to the objective function. This further rules out the application of Newton-type or other gradient-based methods that use traditional finite difference approximations. In addition, if the optimization formulation includes integer variables the complexity grows even further. In this paper we consider three different modeling approaches for a mixed-integer nonlinear optimization problem taken from a set of water resources benchmarking problems. Within this context, we compare the performance of a genetic algorithm, the implicit filtering algorithm, and a branch-and-bound approach that uses sequential surrogate functions. We show that the surrogate approach can greatly improve computational efficiency while locating a comparable, sometimes better, design point than the other approaches.
Article
Full-text available
: This paper presents and analyzes a pattern search method for general constrained optimization based on filter methods for step acceptance. Roughly, a filter method accepts a step that either improves the objective function value or the value of some function that measures the constraint violation. The new algorithm does not compute or approximate any derivatives, penalty constants or Lagrange multipliers. It reduces trivially to the Torczon GPS (generalized pattern search) algorithm when there are no constraints, and indeed, it is formulated here to reduce to the version of GPS designed to handle finitely many linear constraints if they are treated explicitly. A key feature is that it preserves the useful division into search and poll steps. Assuming local smoothness, the algorithm produces a KKT point for a problem related to the original problem. Key words Pattern search algorithm, filter algorithm, surrogate-based optimization, derivative-free convergence analysis, constrained op...
Article
We completed a two-step regional analysis of a coastal groundwater basin to (1) assess regional suitability for managed aquifer recharge (MAR), and (2) quantify the relative impact of MAR activities on groundwater levels and sea water intrusion. The first step comprised an analysis of surface and subsurface hydrologic properties and conditions, using a geographic information system (GIS). Surface and subsurface data coverages were compiled, georeferenced, reclassified, and integrated (including novel approaches for combining related datasets) to derive a spatial distribution of MAR suitability values. In the second step, results from the GIS analysis were used with a regional groundwater model to assess the hydrologic impact of potential MAR placement and operating scenarios. For the region evaluated in this study, the Pajaro Valley Groundwater Basin, California, GIS results suggest that about 7% (15 km(2) ) of the basin may be highly suitable for MAR. Modeling suggests that simulated MAR projects placed near the coast help to reduce sea water intrusion more rapidly, but these projects also result in increased groundwater flows to the ocean. In contrast, projects placed farther inland result in more long-term reduction in sea water intrusion and less groundwater flowing to the ocean. This work shows how combined GIS analysis and modeling can assist with regional water supply planning, including evaluation of options for enhancing groundwater resources.
Article
Particle swarm optimization (PSO) is a population-based, heuristic technique based on social behaviour that performs well on a variety of problems including those with non-convex, non-smooth objective functions with multiple minima. However, the method can be computationally expensive in that a large number of function calls is required. This is a drawback when evaluations depend on an off-the-shelf simulation program, which is often the case in engineering applications. An algorithm is proposed which incorporates surrogates as a stand-in for the expensive objective function, within the PSO framework. Numerical results are presented on standard benchmarking problems and a simulation-based hydrology application to show that this hybrid can improve efficiency. A comparison is made between the application of a global PSO and a standard PSO to the same formulations with surrogates. Finally, data profiles, probability of success, and a measure of the signal-to-noise ratio of the the objective function are used to assess the use of a surrogate.
Article
To size storm-water detention basins, a volume-based method such as that of the Federal Aviation Administration is applicable to small urban catchments. However, the reliability of such a method depends on estimation of the average outflow. In this study, the average outflow from a detention basin is found to be related to the time of concentration of the catchment and the duration of the design storm. This modification eliminates judgment calls and warrants a consistency in volumes and flow rates between the design rainfall hyeotograph and the inflow and outflow hydrographs.
Article
Detention basins are the most popular structural measure for urban food control and have proven effective for both water quantity and quality management. Integrated, watershed-level planning of the layour and sizing of detention systems is essential because localized solutions may actually aggravate the negative impacts of urban drainage. Successive reaching dynamic programming (SRDP) is applied to minimize detention system costs of maintaining ranges of desired downstream peak flow attenuation, with basin and channel routing imbedded within the algorithm. A multiobjective genetic algorithm (MOGA) is also applied to generating non-dominanted solutions for system cost and detention effect for a watershed-level detention system. These algorithms are applied to the layout and design of a stormwater detention system in the Pazam watershed located in southern Taiwan. The case study confirms the robustness and computational efficiency of the SRDP algorith, and teh MOGA generates a wide range of nondiminated solutions for trade-off analysis using a proposed elite solutions conservation procedure.
Article
A model for least cost design of regional stormwater detention networks is developed to aid in identifying cost-effective regional detention basin sites within watersheds undergoing urban development. The model, through the application of dynamic programming techniques, is designed to minimize the cost of required facilities while meeting the hydrologic and hydraulic constraints of the watershed. For this study, model inputs include storage-elevation relationships at each candidate detention basin site, maximum permissible flow in each channel reach, land and dam construction cost functions, watershed hydrologic routing parameters, and design storm hydrograph parameters. The model can be applied to watersheds with multiple branching and existing detention basins. The program is applied to a 16.5 sq mi (42.7 km2) watershed in Austin, Texas. Eight existing and five candidate basins are considered.
Article
Artificial recharge of groundwater is an increasingly important method for augmenting groundwater supply and can have a positive or negative influence on the quality of water resources. We instrumented a managed aquifer recharge (MAR) pond in central coastal California to assess how patterns of infiltration and recharge affect the load of nitrate delivered to the underlying aquifer. The concentration of nitrate in infiltrating water consistently decreased during passage through the first metre of subsurface soils. Enrichment of 18O and 15 N in the residual nitrate in infiltrating water proceeded in a ratio of 1:2, indicating that denitrification plays a significant role in the quantitative reduction of nutrients exported during infiltration through shallow soils. The extent and rate of nitrate removal was spatially and temporally variable across the bottom of the recharge pond, with 30% to 60% of the nitrate load being removed over the first 6 weeks of managed aquifer recharge operation. During the period of highest N loading to the system, when the average infiltration rate was > 1 m/day, the recharge pond achieved a load reduction efficiency of 7 kg NO3−-N/day/ha, which compares favourably to nitrate load reductions achieved by treatment wetlands. Groundwater mounding and water composition below the recharge pond suggest that recharge and subsequent lateral transport occur heterogeneously in the underlying aquifer. Nitrate concentrations in the aquifer following infiltration were lowered primarily by dilution, with little evidence for additional denitrification occurring in the aquifer in comparison to high rates documented during shallow infiltration. Copyright © 2011 John Wiley & Sons, Ltd.
Article
The proliferation of watershed databases in raster Geographic Information System (GIS) format and the availability of radar-estimated rainfall data foster rapid developments in raster-based surface runoff simulations. The two-dimensional physically-based rainfall-runoff model CASC2D simulates spatially-varied surface runoff while fully utilizing raster GIS and radar-rainfall data. The model uses the Green and Ampt infiltration method, and the diffusive wave formulation for overland and channel flow routing enables overbank flow storage and routing. CASC2D offers unique color capabilities to display the spatio-temporal variability of rainfall, cumulative infiltrated depth, and surface water depth as thunderstorms unfold. The model has been calibrated and independently verified to provide accurate simulations of catchment response to moving rainstorms on watersheds with spatially-varied infiltration. The model can accurately simulate surface runoff from flashfloods caused by intense thunderstorms moving across partial areas of a watershed.
Article
Management decisions involving groundwater supply and remediation often rely on optimization techniques to determine an effective strategy. We introduce several derivative-free sampling methods for solving constrained optimization problems that have not yet been considered in this field, and we include a genetic algorithm for completeness. Two well-documented community problems are used for illustration purposes: a groundwater supply problem and a hydraulic capture problem. The community problems were found to be challenging applications due to the objective functions being nonsmooth, nonlinear, and having many local minima. Because the results were found to be sensitive to initial iterates for some methods, guidance is provided in selecting initial iterates for these problems that improve the likelihood of achieving significant reductions in the objective function to be minimized. In addition, we suggest some potentially fruitful areas for future research.
Article
. In this paper we describe and analyze an algorithm for certain box constrained optimization problems that may have several local minima. A paradigm for these problems is one in which the function to be minimized is the sum of a simple function, such as a convex quadratic, and high frequency, low amplitude terms which cause local minima away from the global minimum of the simple function. Our method is gradient based and therefore the performance can be improved by use of quasi-Newton methods. Key words. filtering, projected gradient algorithm, quasi-Newton method AMS(MOS) subject classifications. 65H10, 65K05, 65K10 1. Introduction. In this paper we describe and analyze an algorithm for bound constrained optimization problems that may have several local minima. The type of problem we have in mind is one in which the function to be minimized is the sum of a simple function, such as a convex quadratic, and high frequency, low amplitude terms which cause the local minima. Of particul...
Article
Multi-objective evolutionary algorithms (MOEAs) that use non-dominated sorting and sharing have been criticized mainly for: (1) their O(MN3) computational complexity (where M is the number of objectives and N is the population size); (2) their non-elitism approach; and (3) the need to specify a sharing parameter. In this paper, we suggest a non-dominated sorting-based MOEA, called NSGA-II (Non-dominated Sorting Genetic Algorithm II), which alleviates all of the above three difficulties. Specifically, a fast non-dominated sorting approach with O(MN2) computational complexity is presented. Also, a selection operator is presented that creates a mating pool by combining the parent and offspring populations and selecting the best N solutions (with respect to fitness and spread). Simulation results on difficult test problems show that NSGA-II is able, for most problems, to find a much better spread of solutions and better convergence near the true Pareto-optimal front compared to the Pareto-archived evolution strategy and the strength-Pareto evolutionary algorithm - two other elitist MOEAs that pay special attention to creating a diverse Pareto-optimal front. Moreover, we modify the definition of dominance in order to solve constrained multi-objective problems efficiently. Simulation results of the constrained NSGA-II on a number of test problems, including a five-objective, seven-constraint nonlinear problem, are compared with another constrained multi-objective optimizer, and the much better performance of NSGA-II is observed
Constraint handling for derivative-free optimization
  • A Almomani
A hybrid genetic algorithm with implicit filtering for mixed-integer optimization problems
  • B Ritz
Pajaro valley water management
  • A L Baird
  • S Bovard
  • B Dutta
  • Ji Y Zheng
  • M Farthing
  • E Jenkins