pith. sign in

arxiv: 2501.10842 · v6 · submitted 2025-01-18 · 📡 eess.SY · cs.SY

BOOST: Microgrid Sizing using Ordinal Optimization

Mohamad Chehade , Sami Karaki This is my paper

Pith reviewed 2026-05-23 05:17 UTC · model grok-4.3

classification 📡 eess.SY cs.SY
keywords microgrid sizingordinal optimizationmixed-integer linear programmingphotovoltaic batteryenergy dispatch optimizationsynthetic benchmarkruntime reduction
0
0 comments X

The pith

Ordinal optimization screens microgrid designs with a linear model then re-evaluates the top ones with MILP to recover the optimum at lower cost.

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

The paper tries to solve the coupled design-and-operation problem for residential microgrids where PV and battery sizes must be chosen to reflect realistic dispatch over time. It introduces BOOST, which applies ordinal optimization to rank a large pool of candidate designs using a fast linear program and then runs the slower but accurate MILP model only on the most promising subset. On refreshed synthetic yearly datasets the method recovers the same global optimum found by exhaustive search while reducing runtime by 51.8 percent, and the LP and MILP rankings match exactly on a 10-by-10 grid. A sympathetic reader would care because exhaustive accurate optimization is too slow for routine planning, and the screening step makes the accurate model practical without sacrificing the best design.

Core claim

BOOST combines ordinal optimization with mixed-integer linear programming to size residential microgrids. It screens N candidate battery and PV pairs with a linear programming approximation, then re-evaluates only the top s designs with the full MILP that captures diesel commitment logic. On the base synthetic dataset the best accurate design is 500 kWh battery with 1833.3 kW PV at 13.169 c/kWh; the paper-style choice N=90 and s=18 recovers this global optimum, the LP-MILP Spearman correlation is 1.000 across the 10x10 grid, and the workflow reduces runtime by 51.8 percent relative to exhaustive accurate evaluation. The same pattern holds across the five synthetic configurations, which serve

What carries the argument

Ordinal optimization screening step that selects the top s designs from N candidates according to LP rankings for subsequent accurate MILP evaluation.

If this is right

  • N=90 and s=18 recovers the global accurate optimum on the base synthetic dataset.
  • Runtime is reduced by 51.8 percent relative to exhaustive accurate MILP evaluation.
  • BOOST improves upon dynamic programming and greedy baselines across the tested configurations.
  • LP and MILP rankings are identical (rho=1.000) on the 10x10 design grid.
  • Performance holds on the five synthetic dataset variants used as methodological stress tests.

Where Pith is reading between the lines

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

  • The same screening logic could be tested on other coupled design-operation problems that admit a cheap surrogate model whose ranking is close to the expensive model.
  • If the perfect LP-MILP alignment persists on measured load and weather traces, the runtime saving would allow repeated re-optimization for changing tariffs or degradation.
  • The method's value would increase if the diesel-commitment detail captured by MILP is the dominant source of ranking error in real microgrids.
  • Extending the approach to stochastic or multi-objective versions would require checking whether the surrogate still preserves the correct ordering.

Load-bearing premise

The linear programming approximation produces a ranking of designs that is sufficiently aligned with the ranking produced by the full MILP model.

What would settle it

A new dataset or configuration in which the Spearman rank correlation between LP and MILP objective values drops well below 1 and the designs chosen by the OO step do not include the true MILP optimum.

Figures

Figures reproduced from arXiv: 2501.10842 by Mohamad Chehade, Sami Karaki.

Figure 1
Figure 1. Figure 1: The 5 main components of the system. system size. Unlike the greedy algorithm in [6], the MILP ap￾proach ensures optimality by explicitly considering the entire operating horizon, thus providing better utilization of resources and lower operational costs. This improvement makes BOOST a more robust and effective framework for microgrid sizing and operation. In brief, our main contributions are: 1) We propos… view at source ↗
Figure 2
Figure 2. Figure 2: An overview of BOOST. AP(k) = P rob(|G ∩ S| ≥ k) = min X (g,s) i=k [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
read the original abstract

Sizing a residential microgrid efficiently requires solving a coupled design-and-operation problem: photovoltaic (PV) and battery capacities should be chosen in a way that reflects how the system will actually be dispatched over time. This paper proposes BOOST, or Battery-solar Ordinal Optimization Sizing Technique, which combines ordinal optimization (OO) with mixed-integer linear programming (MILP). OO is used to screen a large set of candidate battery/PV designs with a simple linear model and then re-evaluate only the most promising designs with a more accurate MILP that captures diesel commitment logic. Relative to the original short paper, this expanded manuscript retains the full methodological narrative but refreshes the quantitative section using a new synthetic benchmark dataset suite generated from the released clean reimplementation. The suite contains five yearly synthetic datasets/configurations: base, cheap battery, cheap PV, expensive diesel, and high peak tariff. On the base synthetic dataset, the best accurate design is a 500 kWh battery with 1833.3 kW of PV, achieving 13.169 c/kWh, while BOOST improves upon dynamic programming and greedy baselines. Across the full 10 x 10 design grid, the LP and MILP rankings are effectively identical (rho = 1.000), the paper-style choice of N = 90 and s = 18 recovers the global accurate optimum, and the OO-based workflow reduces runtime by 51.8% relative to exhaustive accurate evaluation on the refreshed synthetic benchmark run. Because these added datasets are synthetic, they should be read as methodological stress tests rather than as direct empirical claims about any specific real-world site. Code is available at https://github.com/MFHChehade/Microgrid-Optimization.

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

0 major / 1 minor

Summary. The manuscript introduces BOOST, which combines ordinal optimization with MILP for residential microgrid sizing: a linear model screens a large candidate set of PV/battery designs, after which only the top-ranked designs are re-evaluated with an accurate MILP that includes diesel commitment. On five synthetic yearly benchmark datasets the method recovers the global optimum (500 kWh battery, 1833.3 kW PV at 13.169 c/kWh) when N=90 and s=18 are used, reports rank correlation rho=1.000 between LP and MILP on the 10x10 grid, reduces runtime by 51.8% versus exhaustive accurate evaluation, and outperforms dynamic-programming and greedy baselines. Code is released.

Significance. The work supplies a reproducible algorithmic workflow that demonstrably reduces the computational burden of coupled design-and-operation optimization while preserving the global optimum on the supplied synthetic benchmarks. Release of the clean reimplementation and explicit verification that the chosen screening parameters recover the true optimum are concrete strengths that support adoption as a methodological tool in energy-systems sizing.

minor comments (1)
  1. [Abstract] Abstract and §1 could add one sentence clarifying that the five datasets are generated synthetically solely for algorithmic stress-testing and do not constitute site-specific empirical validation.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive review and the recommendation to accept. The referee's summary correctly identifies the core contribution of BOOST as a reproducible screening workflow that reduces runtime while recovering the global optimum on the provided synthetic benchmarks.

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper presents an algorithmic workflow (BOOST) that screens candidate microgrid designs via a linear programming approximation and then re-evaluates a shortlist with a full MILP model. Performance is quantified by direct comparison against exhaustive accurate evaluation on externally generated synthetic benchmark datasets; the reported rho = 1.000 and recovery of the global optimum with N=90, s=18 are empirical verification results on those benchmarks rather than any self-referential equation or fitted parameter renamed as a prediction. No derivation step reduces to its own inputs by construction, no uniqueness theorem is imported from prior self-citations, and the central claim remains an externally testable procedure whose correctness is measured against independent exhaustive search.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The claim rests on the domain assumption that the simple LP model preserves ordering relative to the MILP model and on standard mathematical assumptions that MILP solvers return optimal solutions for the dispatch subproblems; no free parameters or invented entities are introduced.

axioms (2)
  • domain assumption The approximate linear model ranking correlates with the accurate MILP ranking
    Invoked in the ordinal optimization screening step that selects which designs to re-evaluate.
  • standard math MILP solvers return globally optimal dispatch solutions for each fixed design
    Required for the accurate evaluation stage to be treated as ground truth.

pith-pipeline@v0.9.0 · 5842 in / 1319 out tokens · 46035 ms · 2026-05-23T05:17:50.583611+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

18 extracted references · 18 canonical work pages

  1. [1]

    Finale COP21,

    United Nations Framework Convention on Climate Change (UNFCCC), “Finale COP21,” 2015, accessed: 2024-11-18. [Online]. Available: https://unfccc.int/news/finale-cop21

    work page 2015

  2. [2]

    Optimal sizing of an autonomous photovoltaic/wind/battery/diesel generator microgrid using grasshopper optimization algorithm,

    A. L. Bukar, C. W. Tan, and K. Y . Lau, “Optimal sizing of an autonomous photovoltaic/wind/battery/diesel generator microgrid using grasshopper optimization algorithm,” Solar Energy , vol. 188, pp. 685– 696, 2019

    work page 2019

  3. [3]

    The potential of hybrid microgrids for mitigating power outage in lebanon,

    R. Chedid and R. Ghajar, “The potential of hybrid microgrids for mitigating power outage in lebanon,” International Journal of Energy and Power Engineering , vol. 13, no. 5, pp. 304–309, 2019

    work page 2019

  4. [4]

    Design and management of hybrid renewable energy system using retscreen software: A case study,

    F. Zaro and N. A. Ayyash, “Design and management of hybrid renewable energy system using retscreen software: A case study,” International Journal of Electrical Engineering and Computer Science , vol. 5, pp. 164–170, 2023

    work page 2023

  5. [5]

    Sizing, optimization, control and energy management of hybrid renewable energy system—a review,

    C. Ammari, D. Belatrache, B. Touhami, and S. Makhloufi, “Sizing, optimization, control and energy management of hybrid renewable energy system—a review,” Energy and Built Environment , vol. 3, no. 4, pp. 399–411, 2022

    work page 2022

  6. [6]

    Optimal design of hybrid renewable energy systems in lebanon,

    A. Richa and S. H. Karaki, “Optimal design of hybrid renewable energy systems in lebanon,” in 2018 IEEE International Multidisciplinary Conference on Engineering Technology (IMCET) . IEEE, 2018, pp. 1–6

    work page 2018

  7. [7]

    A milp algorithm for the optimal sizing of an off-grid hybrid renewable energy system in south tyrol,

    J. C. Alberizzi, M. Rossi, and M. Renzi, “A milp algorithm for the optimal sizing of an off-grid hybrid renewable energy system in south tyrol,” Energy Reports, vol. 6, pp. 21–26, 2020

    work page 2020

  8. [8]

    An enhanced optimal pv and battery sizing model for zero energy buildings considering environmental impacts,

    M. Mehrtash, F. Capitanescu, P. K. Heiselberg, T. Gibon, and A. Bertrand, “An enhanced optimal pv and battery sizing model for zero energy buildings considering environmental impacts,” IEEE Trans- actions on Industry Applications , vol. 56, no. 6, pp. 6846–6856, 2020

    work page 2020

  9. [9]

    An efficient mixed- integer linear programming model for optimal sizing of battery energy storage in smart sustainable buildings,

    M. Mehrtash, F. Capitanescu, and P. K. Heiselberg, “An efficient mixed- integer linear programming model for optimal sizing of battery energy storage in smart sustainable buildings,” in 2020 IEEE Texas Power and Energy Conference (TPEC) . IEEE, 2020, pp. 1–6

    work page 2020

  10. [10]

    Optimal placement and sizing of photo- voltaics and battery storage in distribution networks,

    R. Chedid and A. Sawwas, “Optimal placement and sizing of photo- voltaics and battery storage in distribution networks,” Energy Storage , vol. 1, no. 4, p. e46, 2019

    work page 2019

  11. [11]

    Control, analysis and optimization of hybrid pv-diesel-battery systems for isolated rural city in algeria,

    A. Yahiaoui, K. Benmansour, and M. Tadjine, “Control, analysis and optimization of hybrid pv-diesel-battery systems for isolated rural city in algeria,” Solar Energy, vol. 137, pp. 1–10, 2016

    work page 2016

  12. [12]

    Ordinal optimisation approach for locating and sizing of distributed generation,

    R. Jabr and B. Pal, “Ordinal optimisation approach for locating and sizing of distributed generation,” IET generation, transmission & distri- bution, vol. 3, no. 8, pp. 713–723, 2009

    work page 2009

  13. [13]

    Meter placement for distribution system state estimation: An ordinal optimization approach,

    R. Singh, B. C. Pal, R. A. Jabr, and R. B. Vinter, “Meter placement for distribution system state estimation: An ordinal optimization approach,” IEEE Transactions on Power Systems , vol. 26, no. 4, pp. 2328–2335, 2011

    work page 2011

  14. [14]

    An ordinal optimization based method for power distribu- tion system control,

    E. Atmaca, “An ordinal optimization based method for power distribu- tion system control,” Electric Power Systems Research , vol. 78, no. 4, pp. 694–702, 2008

    work page 2008

  15. [15]

    Near optimal control of fuel cell hybrid electric vehicles in real-time,

    C. Majed, S. H. Karaki, R. Jabr, and F. Panik, “Near optimal control of fuel cell hybrid electric vehicles in real-time,” SAE Technical Paper, Tech. Rep., 2016

    work page 2016

  16. [16]

    Fuel cell hybrid electric vehicle sizing using ordinal optimization,

    S. H. Karaki, R. Dinnawi, R. Jabr, R. Chedid, and F. Panik, “Fuel cell hybrid electric vehicle sizing using ordinal optimization,” SAE Inter- national Journal of Passenger Cars-Electronic and Electrical Systems , vol. 8, no. 2015-01-0155, pp. 60–69, 2015

    work page 2015

  17. [17]

    Analysis of variable reverse osmosis operation powered by solar energy,

    A. Zein, S. Karaki, and M. Al-Hindi, “Analysis of variable reverse osmosis operation powered by solar energy,” Renewable Energy , vol. 208, pp. 385–398, 2023

    work page 2023

  18. [18]

    JA Solar Datasheet,

    JA Solar, “JA Solar Datasheet,” 2021, accessed: [Insert Access Date]. [Online]. Available: https://www.jasolar.com/uploadfile/2021/ 0706/20210706053524693.pdf

    work page 2021