Curses, Tradeoffs, and Scalable Management: Advancing Evolutionary Multiobjective Direct Policy Search to Improve Water Reservoir Operations

Giuliani, Matteo; Castelletti, Andrea; Pianosi, Francesca; Mason, Emanuele; Reed, Patrick M.

doi:10.1061/(asce)wr.1943-5452.0000570

Cited by 198 publications

(183 citation statements)

References 77 publications

(4 reference statements)

Supporting

Mentioning

174

Contrasting

Order By: Relevance

“…It uses an “intelligent search” of the best operating rules, able to obtain them without long trial‐and‐error processes (Celeste & Billib, ; Jacoby & Loucks, ; Johnson, Stedinger, & Staschus, ; Koutsoyiannis & Economou, ; Oliveira & Loucks, ). Its applications have risen considerably in the last decade due to faster heuristic optimization, which can combine simulation and optimization while considering complex performance criteria (Ashbolt, Maheepala, & Perera, ; Giuliani, Castelletti, et al, ; Kumar & Kasthurirengan, ; Lerma, Paredes‐Arquiola, Andreu, & Solera, ; Lerma, Paredes‐Arquiola, Andreu, Solera, & Sechi, ; Shourian, Mousavi, & Tahershamsi, ; Spiliotis et al, ; Yang & Ng, ).…”

Section: A Priori Rule Formsmentioning

confidence: 99%

“…PSO (Figure ) requires establishing a rule form and its parameters. Their optimal values are obtained iteratively combining simulation and optimization (Celeste & Billib, ; Giuliani, Castelletti, et al, ; Koutsoyiannis & Economou, ; Nalbantis & Koutsoyiannis, ). For each iteration, a set of parameter values is chosen and used as input to the simulation algorithm, which obtains the system performance for the given operating rule (rule form plus parameter values) for different inflow scenarios.…”

Section: A Priori Rule Formsmentioning

confidence: 99%

“…PSO approaches can benefit from using general‐purpose decision support systems (DSS) as simulation models (Lerma et al, , ; Shourian et al, ). The combination of MOEAs with PSO, known as evolutionary multiobjective direct policy search (EMODPS) can find Pareto‐optimal operating rules for multipurpose water resource systems (Desreumaux et al, ; Giuliani et al, ; Giuliani, Castelletti, et al, ). Alternatively, optimization methods may be replaced by game theory approaches such as the Nash equilibrium (Wu, Li, Cheng, Miao, & Ying, ).…”

Section: A Priori Rule Formsmentioning

confidence: 99%

See 2 more Smart Citations

Inferring efficient operating rules in multireservoir water resource systems: A review

Macian-Sorribes

Pulido-Velázquez

2019

WIREs Water

View full text Add to dashboard Cite

Coordinated and efficient operation of water resource systems becomes essential to deal with growing demands and uncertain resources in water‐stressed regions. System analysis models and tools help address the complexities of multireservoir systems when defining operating rules. This paper reviews the state of the art in developing operating rules for multireservoir water resource systems, focusing on efficient system operation. This review focuses on how optimal operating rules can be derived and represented. Advantages and drawbacks of each approach are discussed. Major approaches to derive optimal operating rules include direct optimization of reservoir operation, embedding conditional operating rules in simulation‐optimization frameworks, and inferring rules from optimization results. Suggestions on which approach to use depend on context. Parametrization–simulation–optimization or rule inference using heuristics are promising approaches. Increased forecasting capabilities will further benefit the use of model predictive control algorithms to improve system operation. This article is categorized under: Engineering Water > Water, Health, and Sanitation Engineering Water > Methods

show abstract

Section: A Priori Rule Formsmentioning

confidence: 99%

Section: A Priori Rule Formsmentioning

confidence: 99%

Section: A Priori Rule Formsmentioning

confidence: 99%

See 1 more Smart Citation

Inferring efficient operating rules in multireservoir water resource systems: A review

Macian-Sorribes

Pulido-Velázquez

2019

WIREs Water

View full text Add to dashboard Cite

show abstract

“…• sigmodal functions, for which the universal approximation theorem was first proved [54,55]; • radial basis functions, since it has been recently empirically proved that they can be more appropriate in some contexts similar to the one considered here [19].…”

Section: Policy Identificationmentioning

confidence: 99%

NN-Based Implicit Stochastic Optimization of Multi-Reservoir Systems Management

Sangiorgio

Guariso

2018

Water

View full text Add to dashboard Cite

Multi-reservoir systems management is complex because of the uncertainty on future events and the variety of purposes, usually conflicting, of the involved actors. An efficient management of these systems can help improving resource allocation, preventing political crisis and reducing the conflicts between the stakeholders. Bellman stochastic dynamic programming (SDP) is the most famous among the many proposed approaches to solve this optimal control problem. Unfortunately, SDP is affected by the curse of dimensionality: computational effort increases exponentially with the complexity of the considered system (i.e., number of reservoirs), and the problem rapidly becomes intractable. This paper proposes an implicit stochastic optimization approach for the solution of the reservoir management problem. The core idea is using extremely flexible functions, such as artificial neural networks (NN), for designing release rules which approximate the optimal policies obtained by an open-loop approach. These trained NNs can then be used to take decisions in real time. The approach thus requires a sufficiently long series of historical or synthetic inflows, and the definition of a compromise solution to be approximated. This work analyzes with particular emphasis the importance of the information which represents the input of the control laws, investigating the effects of different degrees of completeness. The methodology is applied to the Nile River basin considering the main management objectives (minimization of the irrigation water deficit and maximization of the hydropower production), but can be easily adopted also in other cases.

show abstract

“…However, DP is severely limited by the curse of modeling in designing operating policies conditioned on exogenous information (Tsitsiklis and Van Roy, 1996) and by the curse of multiple objectives in exploring multidimensional tradeoffs (Powell, 2007). We therefore solve Problem (5) by means of an evolutionary multi-objective direct policy search (EMODPS) (Giuliani et al, 2016a), an approximate dynamic programming approach that combines direct policy search, nonlinear approximating networks, and multi-objective evolutionary algorithms. EMODPS allows the direct use of exogenous information through a partially data-driven controller tuning approach (Formentin et al, 2013).…”

Section: Assessment Of the Operational Value Of Virtual Snow Indexesmentioning

confidence: 99%

Using crowdsourced web content for informing water systems operations in snow-dominated catchments

Giuliani

Castelletti

Fedorov

et al. 2016

Hydrol. Earth Syst. Sci.

Self Cite

View full text Add to dashboard Cite

Abstract. Snow is a key component of the hydrologic cycle in many regions of the world. Despite recent advances in environmental monitoring that are making a wide range of data available, continuous snow monitoring systems that can collect data at high spatial and temporal resolution are not well established yet, especially in inaccessible high-latitude or mountainous regions. The unprecedented availability of user-generated data on the web is opening new opportunities for enhancing real-time monitoring and modeling of environmental systems based on data that are public, low-cost, and spatiotemporally dense. In this paper, we contribute a novel crowdsourcing procedure for extracting snow-related information from public web images, either produced by users or generated by touristic webcams. A fully automated process fetches mountain images from multiple sources, identifies the peaks present therein, and estimates virtual snow indexes representing a proxy of the snow-covered area. Our procedure has the potential for complementing traditional snow-related information, minimizing costs and efforts for obtaining the virtual snow indexes and, at the same time, maximizing the portability of the procedure to several locations where such public images are available. The operational value of the obtained virtual snow indexes is assessed for a real-world water-management problem, the regulation of Lake Como, where we use these indexes for informing the daily operations of the lake. Numerical results show that such information is effective in extending the anticipation capacity of the lake operations, ultimately improving the system performance.

show abstract

Curses, Tradeoffs, and Scalable Management: Advancing Evolutionary Multiobjective Direct Policy Search to Improve Water Reservoir Operations

Abstract: General rightsThis document is made available in accordance with publisher policies. Please cite only the published version using the reference above.

Cited by 198 publications

References 77 publications

Inferring efficient operating rules in multireservoir water resource systems: A review

Inferring efficient operating rules in multireservoir water resource systems: A review

NN-Based Implicit Stochastic Optimization of Multi-Reservoir Systems Management

Using crowdsourced web content for informing water systems operations in snow-dominated catchments

Contact Info

Product

Resources

About