pith. machine review for the scientific record. sign in

arxiv: 2605.03374 · v1 · submitted 2026-05-05 · 📡 eess.SY · cs.SY

Recognition: unknown

Event-Based Dynamic Programming for Pumped-Storage Hydropower Scheduling

Bo Yang , Kai Pan , Mohammad Reza Hesamzadeh
Authors on Pith no claims yet

Pith reviewed 2026-05-07 14:10 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords pumped-storage hydropowerevent-based formulationdynamic programmingschedulingmixed-integer programmingreservoir dynamicsbranch-and-boundlinear programming
0
0 comments X

The pith

Pumped-storage hydropower scheduling can be exactly recast as a deterministic dynamic program over sequences of operating-mode events.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes that the standard time-indexed mixed-integer program for single-unit pumped-storage hydropower can be replaced by a representation in which any schedule is a chain of discrete events, each locked to one operating mode such as generation or pumping. Inside each event the best continuous power level is fixed by solving a linear program that respects the mode's limits and reservoir dynamics. This construction yields an equivalent dynamic program on an event network whose states track mode transitions and cumulative energy, preserving all ramping, minimum up-time, and start-up cost constraints. The reformulation is exact and modular, so new modes can be added by defining new event types without rebuilding the network structure. Tractable methods follow from a finite-grid discretization that produces a linear program or from a branch-and-bound procedure that uses linear-program bounds on the continuous-state version.

Core claim

Based on this construction, the original time-indexed mixed-integer formulation is reformulated exactly as a deterministic dynamic program on an event network, in which an operating schedule is represented as a sequence of mode-specific events with dispatch decisions within each event determined by linear programs.

What carries the argument

The event network, which replaces time-indexed binary variables with transitions between mode-specific events whose internal dispatch is optimized by linear programs.

If this is right

  • Additional operating modes such as hydraulic short-circuit operation can be incorporated by adding corresponding event modules without significantly changing the overall event-network structure.
  • A finite-grid approximation of the event network produces a linear programming formulation for the discretized model.
  • An event-based branch-and-bound algorithm that uses linear-program bounds solves the continuous-state version of the problem.
  • Numerical tests show the event-based approach is computationally competitive with the conventional time-indexed formulation while retaining modeling flexibility.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same event-network idea could be applied to other unit-commitment problems that mix discrete mode switches with continuous dispatch decisions.
  • Because the network separates mode logic from continuous optimization, stochastic extensions might be handled by attaching scenario trees only to the event transitions.
  • Optimal policies extracted from the dynamic program might reveal recurring patterns in when and how long a plant should stay in each mode.

Load-bearing premise

Every feasible and optimal operating schedule can be represented without loss as a finite sequence of mode-specific events in which the continuous dispatch decisions are exactly optimal for the linear program solved inside each event.

What would settle it

An optimal schedule obtained from the time-indexed mixed-integer program whose cost is strictly lower than every feasible path through the event network, or a feasible schedule that cannot be expressed as any finite sequence of the defined mode events.

Figures

Figures reproduced from arXiv: 2605.03374 by Bo Yang, Kai Pan, Mohammad Reza Hesamzadeh.

Figure 1
Figure 1. Figure 1: Scalability of LP, MILP, and B&B with respect to the number of view at source ↗
read the original abstract

This paper studies the single-unit pumped-storage hydropower (PSH) plant scheduling problem with reservoir dynamics, generation and pumping limits, ramping constraints, start-up and shut-down costs, and minimum up/down-time requirements. A new event-based formulation is proposed in which an operating schedule is represented as a sequence of mode-specific events, with dispatch decisions within each event determined by linear programs. Based on this construction, the original time-indexed mixed-integer formulation is reformulated exactly as a deterministic dynamic program on an event network. The framework is modular and can be extended to incorporate additional operating modes, such as hydraulic short-circuit operation, by introducing corresponding event modules without significantly changing the overall event-network structure. To obtain tractable solution methods, a finite-grid approximation of the event network is developed, leading to a linear programming formulation for the discretized model. In addition, an event-based branch-and-bound algorithm with linear program-based bounds is proposed for the continuous-state problem. Numerical results demonstrate that the proposed event-based framework provides a computationally effective alternative to the conventional time-indexed formulation, while offering substantial modeling flexibility for PSH scheduling problems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 3 minor

Summary. The paper claims that the standard time-indexed MILP formulation for single-unit pumped-storage hydropower scheduling—with reservoir dynamics, generation/pumping limits, ramping, start-up/shut-down costs, and minimum up/down times—can be exactly reformulated as a deterministic dynamic program on an event network. Operating schedules are represented as finite sequences of mode-specific events, with intra-event continuous dispatch decisions obtained by solving linear programs; modular event modules encode transitions, ramping, and costs. The framework extends to additional modes (e.g., hydraulic short-circuit) without restructuring the network. Tractable methods include a finite-grid discretization yielding an LP and an event-based branch-and-bound algorithm using LP bounds. Numerical results are presented to show computational effectiveness versus the conventional formulation.

Significance. If the exact equivalence holds, the contribution is significant for energy-systems optimization: it replaces a monolithic time-indexed MILP with a modular DP whose state is defined by events rather than time steps, while preserving optimality. The explicit construction from event decomposition (rather than parameter fitting) and the modularity for new operating modes are strengths. Credit is given for supplying both a discretized LP approximation and a continuous-state B&B procedure, together with the claim of an exact (non-heuristic) reformulation.

major comments (2)
  1. [Event-network construction (§3)] Event-network construction (abstract and §3): the central claim of exact equivalence between the original time-indexed MILP and the event-based DP rests on the assertion that every feasible schedule can be represented as a sequence of mode-specific events whose intra-event LPs recover the optimal continuous dispatch. The manuscript must supply an explicit theorem (with proof sketch) showing that the minimum up/down-time and ramping constraints are preserved across variable-duration events without relaxation or additional binaries; otherwise the load-bearing equivalence is not fully established.
  2. [Event-transition equations] Reservoir-dynamics handling (event-transition equations): because events have variable lengths, the continuous-state reservoir level update between events must be shown to match the original time-indexed dynamics exactly for arbitrary start times. If the transition function is only approximated on the finite grid, the exactness claim for the continuous DP requires a separate argument that is currently missing.
minor comments (3)
  1. [Notation] Notation table: introduce a consolidated table listing all event-module parameters, state variables, and transition costs; several symbols (e.g., event duration bounds, mode-specific ramp rates) are introduced inline and are easy to overlook.
  2. [Numerical results] Numerical section: the reported CPU times and objective values should be accompanied by instance sizes (number of time periods, reservoir discretization) and a direct side-by-side comparison table against a standard MILP solver on identical test cases to substantiate the 'computationally effective' statement.
  3. [Figures] Figure clarity: the event-network diagram (presumably Figure 2 or 3) should label the nodes with the corresponding mode and the arcs with the encoded constraints (ramping, min up/down) so that the modularity claim is visually verifiable.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the positive evaluation, the recommendation of minor revision, and the insightful comments on the event-network construction and reservoir dynamics. These points help strengthen the presentation of the exact equivalence. We will revise the manuscript to include an explicit theorem with proof sketch and additional clarifications on the transitions. Below we respond point by point.

read point-by-point responses

  1. Referee: [Event-network construction (§3)] Event-network construction (abstract and §3): the central claim of exact equivalence between the original time-indexed MILP and the event-based DP rests on the assertion that every feasible schedule can be represented as a sequence of mode-specific events whose intra-event LPs recover the optimal continuous dispatch. The manuscript must supply an explicit theorem (with proof sketch) showing that the minimum up/down-time and ramping constraints are preserved across variable-duration events without relaxation or additional binaries; otherwise the load-bearing equivalence is not fully established.

    Authors: We agree that an explicit theorem would improve clarity. In the revised manuscript we will add Theorem 3.1 in Section 3, which asserts exact equivalence between the time-indexed MILP and the event-based DP. The proof sketch will show both directions: every feasible MILP schedule aggregates into a valid event sequence (consecutive identical-mode intervals become single events whose durations satisfy minimum up/down times by construction, with ramping enforced inside events by the intra-event LP and at boundaries by the transition modules); conversely, every feasible event sequence expands to a time-indexed schedule satisfying all original constraints. The event network encodes transitions and minimum durations structurally, so no additional binary variables are required. This addition preserves the claimed exactness without altering the modeling approach. revision: yes

  2. Referee: [Event-transition equations] Reservoir-dynamics handling (event-transition equations): because events have variable lengths, the continuous-state reservoir level update between events must be shown to match the original time-indexed dynamics exactly for arbitrary start times. If the transition function is only approximated on the finite grid, the exactness claim for the continuous DP requires a separate argument that is currently missing.

    Authors: The continuous-state DP employs an exact transition function: the reservoir update is the net volume change over the event duration, computed from the constant dispatch level returned by the intra-event LP. Because the original MILP reservoir balance is linear per time step, aggregation over any duration yields an identical net change for the same start time and dispatch profile. The finite-grid LP approximates only the state space for tractability; the continuous DP and the event-based branch-and-bound retain the precise linear transition. We will insert a clarifying paragraph in Section 3.2 and a remark in Section 4 that explicitly separates the exact continuous transitions from the state discretization, thereby completing the argument for exact equivalence in the continuous case. revision: yes

Circularity Check

0 steps flagged

No significant circularity; reformulation is a direct constructive equivalence

full rationale

The central claim is an exact reformulation of the time-indexed MILP into a deterministic DP on an event network, achieved by representing schedules as finite sequences of mode-specific events whose intra-event continuous dispatch is solved exactly by LPs. This is a modeling construction built from the event decomposition, ramping/min-up/down constraints, and modular event modules; it does not reduce to fitted parameters, self-referential definitions, or load-bearing self-citations. The finite-grid LP and event-based B&B are presented as solution methods for the resulting model rather than alterations to the claimed equivalence. No steps in the derivation chain collapse by construction to the inputs.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The central claim rests on two domain assumptions about event decomposition and linear-program optimality within events; the finite-grid approximation introduces a tunable discretization parameter whose value affects solution quality but is not fitted to data in the abstract.

free parameters (1)
  • discretization grid resolution
    Controls the accuracy of the finite-grid linear-programming approximation of the continuous event network.
axioms (2)
  • domain assumption Any feasible operating schedule can be represented exactly as a sequence of mode-specific events.
    Invoked to convert the original time-indexed MIP into a dynamic program on the event network.
  • domain assumption Optimal dispatch decisions inside each event are obtained by solving a linear program.
    Used to separate the combinatorial event sequencing from the continuous power decisions.

pith-pipeline@v0.9.0 · 5497 in / 1470 out tokens · 124065 ms · 2026-05-07T14:10:08.024333+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

31 extracted references · 1 canonical work pages

  1. [1]

    Optimizing pumped storage hydropower for multiple grid services,

    X. Ma, D. Wu, D. Wang, B. Huang, K. Desomber, T. Fu, and M. Weimar, “Optimizing pumped storage hydropower for multiple grid services,” Journal of Energy Storage, vol. 51, p. 104440, 2022

    2022

  2. [2]

    Pumped stor- age hydropower factsheet,

    International Hydropower Association, “Pumped stor- age hydropower factsheet,” 2021. [Online]. Available: https://www.hydropower.org/factsheets/pumped-storage

    2021

  3. [3]

    Optimal operation scheduling of pumped storage hydro power plant in power system with a large penetration of photovoltaic generation using genetic algorithm,

    R. Aihara, A. Yokoyama, F. Nomiyama, and N. Kosugi, “Optimal operation scheduling of pumped storage hydro power plant in power system with a large penetration of photovoltaic generation using genetic algorithm,” in2011 IEEE Trondheim PowerTech. IEEE, 2011, pp. 1–8

    2011

  4. [4]

    Hydropower special market report,

    International Energy Agency, “Hydropower special market report,”

  5. [5]

    Available: https://www.iea.org/reports/hydropower- special-market-report

    [Online]. Available: https://www.iea.org/reports/hydropower- special-market-report

  6. [6]

    Multistage stochastic power generation scheduling co-optimizing energy and ancillary services,

    J. Huang, K. Pan, and Y . Guan, “Multistage stochastic power generation scheduling co-optimizing energy and ancillary services,”INFORMS Journal on Computing, vol. 33, no. 1, pp. 352–369, 2021

    2021

  7. [7]

    Optimal energy and reserve scheduling of pumped-storage power plants considering hydraulic short-circuit operation,

    M. Chazarra, J. I. P ´erez-D´ıaz, and J. Garc´ıa-Gonz´alez, “Optimal energy and reserve scheduling of pumped-storage power plants considering hydraulic short-circuit operation,”IEEE Transactions on Power Systems, vol. 32, no. 1, pp. 344–353, 2016

    2016

  8. [8]

    Optimal joint energy and secondary regulation reserve hourly scheduling of variable speed pumped storage hydropower plants,

    M. Chazarra, J. I. P ´erez-D´ıaz, and J. Garc ´ıa-Gonz´alez, “Optimal joint energy and secondary regulation reserve hourly scheduling of variable speed pumped storage hydropower plants,”IEEE Transactions on Power Systems, vol. 33, no. 1, pp. 103–115, 2018

    2018

  9. [9]

    Pumped-storage hydro- turbine bidding strategies in a competitive electricity market,

    N. Lu, J. H. Chow, and A. A. Desrochers, “Pumped-storage hydro- turbine bidding strategies in a competitive electricity market,”IEEE transactions on power systems, vol. 19, no. 2, pp. 834–841, 2004

    2004

  10. [10]

    Optimization-based scheduling of hydrothermal power systems with pumped-storage units,

    X. Guan, P. B. Luh, H. Yen, and P. Rogan, “Optimization-based scheduling of hydrothermal power systems with pumped-storage units,” IEEE Transactions on Power Systems, vol. 9, no. 2, pp. 1023–1031, 1994

    1994

  11. [11]

    Optimal short- term dispatch of pumped-storage hydropower plants including hydraulic short circuit,

    F. Gerini, E. Vagnoni, R. Cherkaoui, and M. Paolone, “Optimal short- term dispatch of pumped-storage hydropower plants including hydraulic short circuit,”IEEE Transactions on Power Systems, vol. 40, no. 3, pp. 2050–2060, 2024

    2050

  12. [12]

    A configuration based pumped storage hydro model in the miso day-ahead market,

    B. Huang, Y . Chen, and R. Baldick, “A configuration based pumped storage hydro model in the miso day-ahead market,”IEEE Transactions on Power Systems, vol. 37, no. 1, pp. 132–141, 2022

    2022

  13. [13]

    Approximating input-output curve of pumped storage hydro plant: A disjunctive convex hull method,

    S. Wang, J. Liu, R. Bo, and Y . Chen, “Approximating input-output curve of pumped storage hydro plant: A disjunctive convex hull method,”IEEE Transactions on Power Systems, vol. 38, no. 1, pp. 63–74, 2022

    2022

  14. [14]

    Optimization of pumped hydro energy storage systems under uncertainty: A review,

    P. Toufani, E. C. Karakoyun, E. Nadar, O. B. Fosso, and A. S. Kocaman, “Optimization of pumped hydro energy storage systems under uncertainty: A review,”Journal of Energy Storage, vol. 73, p. 109306, 2023

    2023

  15. [15]

    Secured reserve scheduling of pumped-storage hydropower plants in iso day- ahead market,

    Y . Liu, L. Wu, Y . Yang, Y . Chen, R. Baldick, and R. Bo, “Secured reserve scheduling of pumped-storage hydropower plants in iso day- ahead market,”IEEE Transactions on Power Systems, vol. 36, no. 6, pp. 5722–5733, 2021

    2021

  16. [16]

    Convex hull model for a single-unit commitment problem with pumped hydro storage unit,

    M. Qu, T. Ding, Y . Sun, C. Mu, K. Pan, and M. Shahidehpour, “Convex hull model for a single-unit commitment problem with pumped hydro storage unit,”IEEE Transactions on Power Systems, vol. 38, no. 5, pp. 4867–4878, 2023

    2023

  17. [17]

    Lin- earization method for large-scale hydro-thermal security-constrained unit commitment,

    M. Qu, T. Ding, C. Mu, X. Zhang, K. Pan, and M. Shahidehpour, “Lin- earization method for large-scale hydro-thermal security-constrained unit commitment,”IEEE Transactions on Automation Science and Engineer- ing, vol. 21, no. 2, pp. 1754–1766, 2024

    2024

  18. [18]

    Polynomial time algorithms and ex- tended formulations for unit commitment problems,

    Y . Guan, K. Pan, and K. Zhou, “Polynomial time algorithms and ex- tended formulations for unit commitment problems,”IISE Transactions, vol. 50, no. 8, pp. 735–751, 2018

    2018

  19. [19]

    Con- vex hull for self-scheduling energy-intensive enterprises with demand response regulations,

    Y . Xiao, T. Ding, C. Mu, K. Pan, B. Zhang, and M. Shahidehpour, “Con- vex hull for self-scheduling energy-intensive enterprises with demand response regulations,”IEEE Transactions on Power Systems, vol. 40, no. 3, pp. 2003–2013, 2025

    2003

  20. [20]

    Polyhedral characteri- zation of discrete dynamic programming,

    R. K. Martin, R. L. Rardin, and B. A. Campbell, “Polyhedral characteri- zation of discrete dynamic programming,”Operations Research, vol. 38, no. 1, pp. 127–138, 1990

    1990

  21. [21]

    Arc flow formulations based on dynamic programming: The- oretical foundations and applications,

    V . L. de Lima, C. Alves, F. Clautiaux, M. Iori, and J. M. Val ´erio de Carvalho, “Arc flow formulations based on dynamic programming: The- oretical foundations and applications,”European Journal of Operational Research, vol. 296, no. 1, pp. 3–21, 2022

    2022

  22. [22]

    A poly- hedral study of production ramping,

    P. Damcı-Kurt, S. K ¨uc ¸¨ukyavuz, D. Rajan, and A. Atamt ¨urk, “A poly- hedral study of production ramping,”Mathematical Programming, vol. 158, pp. 175–205, 2016

    2016

  23. [23]

    Block-based state-expanded network models for multi- activity shift scheduling,

    M. R ¨omer, “Block-based state-expanded network models for multi- activity shift scheduling,”Journal of Scheduling, vol. 27, no. 4, pp. 341–361, 2024

    2024

  24. [24]

    A computationally efficient mixed-integer linear formulation for the thermal unit commitment problem,

    M. Carri ´on and J. M. Arroyo, “A computationally efficient mixed-integer linear formulation for the thermal unit commitment problem,”IEEE Transactions on Power Systems, vol. 21, no. 3, pp. 1371–1378, 2006

    2006

  25. [25]

    Tight mixed integer linear programming formulations for the unit commitment problem,

    J. Ostrowski, M. F. Anjos, and A. Vannelli, “Tight mixed integer linear programming formulations for the unit commitment problem,”IEEE Transactions on Power Systems, vol. 27, no. 1, pp. 39–46, 2012

    2012

  26. [26]

    Tight and compact milp formulation of start-up and shut-down ramping in unit commit- ment,

    G. Morales-Espa ˜na, J. M. Latorre, and A. Ramos, “Tight and compact milp formulation of start-up and shut-down ramping in unit commit- ment,”IEEE Transactions on Power Systems, vol. 28, no. 2, pp. 1288– 1296, 2013

    2013

  27. [27]

    A polyhedral study of the integrated minimum- up/-down time and ramping polytope,

    K. Pan and Y . Guan, “A polyhedral study of the integrated minimum- up/-down time and ramping polytope,”arXiv preprint arXiv:1604.02184, 2016

    work page arXiv 2016

  28. [28]

    On mixed-integer pro- gramming formulations for the unit commitment problem,

    B. Knueven, J. Ostrowski, and J.-P. Watson, “On mixed-integer pro- gramming formulations for the unit commitment problem,”INFORMS Journal on Computing, vol. 32, no. 4, pp. 857–876, 2020

    2020

  29. [29]

    Solving storage- integrated long-term unit commitment with guaranteed bounds on sub- optimality,

    S. Feng, W. Wei, Z. Guo, Z. Dong, and S. Mei, “Solving storage- integrated long-term unit commitment with guaranteed bounds on sub- optimality,”IEEE Transactions on Power Systems, 2025

    2025

  30. [30]

    Coupling pumped hydro energy storage with unit commitment,

    K. Bruninx, Y . Dvorkin, E. Delarue, H. Pand ˇzi´c, W. D’haeseleer, and D. S. Kirschen, “Coupling pumped hydro energy storage with unit commitment,”IEEE Transactions on Sustainable Energy, vol. 7, no. 2, pp. 786–796, 2015

    2015

  31. [31]

    Dynamic programming via linear programming,

    ˙I. E. B ¨uy¨uktahtakın, “Dynamic programming via linear programming,” inWiley Encyclopedia of Operations Research and Management Science, J. J. Cochran, L. A. J. Cox, P. Keskinocak, J. P. Kharoufeh, and J. C. Smith, Eds. Hoboken, NJ: John Wiley & Sons, 2011, pp. 1561–1566. 11 APPENDIXI HYDRAULICSHORT-CIRCUITEXTENSION This appendix collects the extensions...

    2011