A multi-stakeholder portfolio optimization framework applied to stormwater best management practice (BMP) selection

Addressing Challenges of Urban Water Management in Chinese Sponge Cities via Nature-Based Solutions

“…1970s, "Best Management Practices" (BMPs), which include nature-based approaches, were introduced in North America to control urban stormwater pollution. BMPs can be structural (i.e., sediment control fences to retain sediment on a construction site) or non-structural (i.e., monitoring to ensure sediment concentrations downstream of construction sites stay below a permitted level) [24]. In the US, the adoption of BMPs is regulated by the Environmental Protection Agency (EPA), which has developed a BMPs Siting Tool to guide urban planners, architects, and developers [25].…”

Introductory overview: Optimization using evolutionary algorithms and other metaheuristics

“…It should be noted that the translation of environmental and water resources management problems into the somewhat rigid formulation outlined above can be challenging, especially for real-world problems, which are often complex, high-dimensional, ill-defined and not well structured. Consequently, where possible, stakeholders should be directly involved in the formulation of optimization problems, informing the objectives, decisions and constraints as part of an interactive process (Di Matteo et al, 2017a;Smith et al, 2017;Wu et al, 2016).…”

“…EAs have proven to be highly effective in solving a wide range of environmental problems , as they can be used in realistic decision-contexts. This is because they (i) can be used as one element in broader, participatory environmental decision-making processes (Di Matteo et al, 2017a;Kaim et al, 2018;Piscopo et al, 2015;Wu et al, 2016), (ii) can be linked with existing simulation models to assist with the exploration of large solution spaces (Maier et al, 2015), (iii) can cater to multiple competing objectives (Newland et al, 2018;Verstegen et al, 2017), (iv) can take into account uncertainties (Beh et al, 2017;Eker and Kwakkel, 2018) and (v) generally provide a number of "good" solutions that can be explored further, rather than being prescriptive (Di Matteo et al, 2017a;Kaim et al, 2018). Their main disadvantages are (i) that they are potentially computationally expensive, although this primarily depends on the computational efficiency of the simulation model with which they are linked, (ii) that they are not guaranteed to identify the globally optimal solution (from a mathematical perspective) and (iii) that they generally need to be tuned to the problem under consideration (Maier et al, 2014;Mala-Jetmarova et al, 2017).…”

“…This enables consideration of factors other than those captured in the mathematical formulation of the optimization problem when selecting the final "optimal" solution. As a result, decision-makers have greater control in terms of using their judgement and intuition to select the final solution based on a number of good solutions "suggested" by the optimization algorithm (Di Matteo et al, 2017b). In this way, EAs are used to assist with "sifting through" the very large solution spaces that are a feature of environmental problems (see Table 1) in order to identify a set of near-optimal candidate solutions that can then be scrutinized by decision-makers to identify those that make most sense, either informally or with the aid of multicriteria decision analysis methods (e.g.…”

Urban Stormwater Characterization, Control and Treatment

“…Other authors expanded the set of optimization functions considered, to include avoided flood damage costs (Cunha et al, 2017) as well as benefit intensity and benefit profile associated with a broader range of societal benefits, or ecosystem services, potentially provided by LID practices (Fenner, 2017). Di Matteo et al (2017b) presented one of the broadest optimization frameworks, which sought to identify optimal solutions across multiple stakeholders. They applied this framework to a case watershed in Australia to demonstrate how costs (here, installation and maintenance costs of LID practices) and benefits (here, stormwater harvesting, nitrogen load reduction and amenity improvement) could be optimized so that they were spread evenly over four different stakeholder groups.…”