Pith. sign in

REVIEW 3 major objections 5 minor 86 references

A Nurse Staffing and Scheduling Problem with Bounded Flexibility and Demand Uncertainty

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Multi-stage nurse staffing under uncertain demand is exactly a two-stage stochastic program via block-separable recourse, and a generative scheduler plus bounded flexibility finds near-optimal rosters with large savings over deterministic…

desk verdict A careful applied model undermined by an unproven—and likely false—block-separable recourse claim that carries the entire two-stage reformulation and all computational results. read the letter →

arxiv 2505.22124 v2 pith:7UT7NXNL submitted 2025-05-28 math.OC

Si Zhang , Paul Mingzheng Tang , Hoong Chuin Lau This is my paper
classification math.OC MSC 90C1590C1190B35
keywords nursestaffingschedulingmulti-stagestochasticprogrammingblock-separablerecourseboundedflexibilitygenerativeflownetworksdemanduncertaintyworkregularitypolicy
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper attacks a practical tension in hospital wards: strict rostering rules, especially a time-regularity policy fixing how many day, evening, and night slots a nurse may work across the horizon, versus nurses' desire for flexible, requested shifts. It introduces bounded flexibility as the compromise and models the staffing-and-scheduling problem as a multi-stage stochastic program in which staffing levels are set before each stage's demand is revealed and detailed rosters adjust afterward. The central structural claim is that this multi-stage program has block-separable recourse, so it is exactly equivalent to a two-stage stochastic program, a reduction that makes large instances tractable. On anonymized data from a Singapore hospital ward, the stochastic model beats expected-value planning in every tested case, and a modest reduction in regularity (from 0.6 to 0.4 in the share of regular-policy nurses) substantially raises request satisfaction.

What carries the argument

The load-bearing object is block-separable recourse in the Birge-Louveaux sense: within each stage, decisions split into aggregate staffing decisions (nurse counts, set before demand is seen) and detailed scheduling decisions (shift assignments, made after), with a block-diagonal recourse matrix so consecutive stages interact only through headcount variables. Appendix I verifies the four Birge-Louveaux conditions, and Appendix J uses them to collapse the scenario tree into a first stage of aggregate staffing decisions plus independent per-stage detailed recourse problems, the reformulation that makes a commercial MILP solver or the generative heuristic practical. The second mechanism is the Generative Flow Network trained with trajectory-balance loss, which samples candidate rosters with probability proportional to their reward and thereby supplies the diverse candidate set; the third is the work-policy variables p1, p2, and p3 that encode bounded flexibility as a cap on how many time slots a nurse may be assigned across the planning horizon.

What would settle it

Solve one small instance, such as two stages with two demand realizations each and a handful of nurses, twice: once as the full multi-stage model (F3) and once as the reduced two-stage model (G1), both to proven optimality with identical costs and data. Any instance where the two optimal values differ disproves the without-loss-of-accuracy claim; a cheaper structural check is to test whether constraint E9, whose work-policy variables are fixed at stage 0 and appear in every stage's scheduling constraints, admits a truly block-diagonal recourse matrix.

Watch

Extended reading notes

Core claim

The paper's central claim, stated for a fair reader: the multi-stage nurse staffing and scheduling problem with stage-by-stage demand revelation, bounded flexibility, and a full hospital roster-rule set is exactly equivalent (Theorem 1 and Remark 1) to a two-stage stochastic program with recourse, because each stage's detailed scheduling decisions couple to the rest of the problem only through the aggregate nurse count. With that reduction, the paper builds a generative flow network scheduler (GNS) that samples diverse rosters in proportion to their reward and, after demand realizes, quickly evaluates the candidates to pick the best. On real ward data the paper reports that the stochastic model outperforms expected-value planning in every case (the value of the stochastic solution is always positive), that the multi-stage formulation beats a two-stage simplification, that the generative scheduler reaches negative gaps versus the commercial solver on several 40-nurse instances, and that flexibility rises from around 0.3 to 0.8 as the regularity level falls, with the steepest gains coming from a small relaxation at the high-regularity end.

Load-bearing premise

The reduction's correctness rests on the claim that each stage's detailed shift assignments interact with the rest of the plan only through the total nurse count; if the shared stage-0 work-policy indicators couple detailed decisions across stages and scenarios, the recourse matrix is not block-diagonal and the two-stage reformulation is no longer exact.

Editorial extensions

If this is right

  • The multi-stage program can be solved through the two-stage reduction, and the positive value-of-stochastic-solution in all tested cases means planning across demand scenarios instead of on expected demand cuts total cost.
  • The value of the stochastic solution grows with ward size, so uncertainty-aware staffing matters more in larger wards with more nurses.
  • A modest drop in the regularity level (from 0.6 to 0.4 in the share of regular-policy nurses) substantially raises flexibility, suggesting hospitals can trade a little schedule regularity for markedly better request satisfaction.
  • The generative approach produces a diverse set of candidate rosters with high-reward schedules sampled most often, and on 40-nurse instances it matches or beats the commercial solver within the two-hour limit, with gaps down to -9.28 percent.
  • The demand coverage gap in the real ward can be held within [-2, 1] per time slot while balancing mismatch across slots, so the model captures supply-demand balance better than the manual current practice.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the block-separable reduction is exact, the same trick should transfer to other staged workforce problems, such as call-center staffing, emergency-department coverage, or flight-crew rostering, wherever the only cross-stage coupling is headcount.
  • The reduction's validity deserves a stress test on the shared stage-0 work-policy indicators in constraint E9: if a hospital later adds cross-stage pattern rules, such as a night shift in one stage forbidding a morning shift in the next, the recourse matrix gains off-diagonal blocks and the no-loss-of-accuracy claim would need to be re-proven.
  • The GNS reward is computed on a single expected-demand scenario and robustness comes from sampling diversity; a natural testable extension is to make the reward itself a small multi-scenario evaluation, which could close the gaps observed in low-demand cases where GNS trails the stochastic program.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper studies an integrated nurse staffing and scheduling problem with demand uncertainty and a hospital-level time-regularity policy, introducing the notion of bounded flexibility. It formulates a deterministic MILP (Model 2), a multi-stage stochastic program with aggregate and detailed decisions (Model 3/4), and claims that the latter has block-separable recourse, allowing an equivalent two-stage reformulation (Model 5) without loss of accuracy. A Generative Flow Network (GFN) based heuristic, GNS, is proposed to generate diverse candidate schedules, and experiments on anonymized data from a Singapore hospital are reported, including comparisons against a commercial solver and a value-of-stochastic-solution analysis.

Significance. If the block-separable reduction were valid, the paper would offer a practically useful way to solve a realistic multi-stage nurse staffing and scheduling problem and would demonstrate the value of generative flow networks in this domain. The paper is also valuable for its detailed deterministic model, the explicit treatment of work-policy regularity as bounded flexibility, and the use of real hospital data. These modeling contributions are real and could support future work. However, the central structural claim on which the computational story rests is not established, and the reported numerical comparisons are weakened by large optimality gaps on the stochastic model.

major comments (3)
  1. [Appendix I, Theorem 1, with constraint (E9)] The proof of Theorem 1 asserts that the Birge-Louveaux block-separable conditions I1-I4 hold, but it does not verify them against constraints E7-E19. In particular, constraint (E9), y^j_i >= sum_{s in S_i cap S_j} x^d,s_{i,h}(omega_h) for all h, links the stage-0 detailed-level work-policy indicator y^j_i to every stage-h detailed assignment. In the block-separable form (I3)-(I4), this creates a nonzero S(omega_h) block in the technology matrix, violating condition I2/I3(b) and making the recourse matrix non-block-diagonal. The claim that detailed decisions in one stage have no direct effect on decisions in other stages is therefore false for this model.
  2. [Appendix J, Equations (J4)-(J6), and objective (F5)] The decomposition of the recourse function into Q^A_h and Q^D_h silently drops the adjustment cost CA * U^d,s_{i,h} and the constraints (F1)-(F2) that define U relative to the stage-0 detailed schedule x^d,s_i. In the actual recourse function (F5), every stage-h problem depends on the first-stage detailed schedule x through both the objective and the constraints. Thus Q^D_h as defined in (J6) is not the actual detailed recourse, and the equivalence stated in Remark 1 is not proven. Model 5 is merely an extensive-form rewriting of Model 4, not a two-stage reduction that removes couplings.
  3. [Table II and Section V.G] Many reported SP optimality gaps are extremely large (for example, cases 40-0.01 to 40-0.25 show SP gaps of 488-519%, and 30-0.01 shows 56.60%). The stochastic objective values used in the cost-savings claims and in the negative GNS gaps are therefore best-known-feasible values, not optima or near-optima. The value-of-stochastic-solution results in Table III inherit this issue, so the conclusion that the stochastic model achieves significant cost savings is not rigorously supported unless instances are solved to optimality or valid lower bounds are provided.
minor comments (5)
  1. [Section IV.A] The sentence 'According to Proposition 1' refers to a proposition that is not labeled; it should refer to Theorem 1.
  2. [Appendix H] The notation table lists y^j_i only as a stage-0 variable implicitly through chi; it should explicitly state that y^j_i is a stage-0 decision variable that appears in stage-h constraints, since this is central to the structure.
  3. [Table II] The two GAP columns are not comparable: the SP gap is the solver's branch-and-cut gap, while the GNS gap is a relative difference to the SP objective value. The table caption should define each gap and note that a negative GNS gap means GNS beat the solver's feasible solution, not the true optimum.
  4. [Appendix J, expansion (J7)] The recursion in (J7) omits the adjustment term and the coupling constraints in every stage; the notation Q^D_h is used before it is defined in (J6), and the indices h and |H| are inconsistent between the recursion and the expansion.
  5. [Section V.G] The definition of VSS uses U*_5, the optimal value of Model 5, but if the equivalence of Model 5 to the multi-stage problem is not established, the comparison between PP and TP needs to be reinterpreted.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the derivation chain rests on external Birge–Louveaux theory and a legitimate surrogate reward; the main proof gap is a correctness issue, not an input–output identity.

full rationale

The paper's central computations are not circular. The block-separable reduction (Theorem 1, Remark 1) invokes the external textbook framework of Birge and Louveaux [9], and no load-bearing step is justified by author self-citation. The GNS reward function (Eq. 9) is explicitly a single-scenario simplification of the model objective used to train a sampler; evaluating the resulting candidate schedules under the stochastic model is a standard optimization/evaluation loop, not a fitted parameter being relabeled as a prediction. The VSS comparisons in Section V.G are defined against genuinely different benchmark models. The only serious concern is that Appendix I verifies the Birge–Louveaux block-separable conditions by assertion, and Appendix J's decomposition (J4–J6) drops the adjustment term that appears in the original recourse F5, while Constraint E9 links the stage-0 detailed variable y_j^i to every stage and scenario. If sustained, this would invalidate the claimed equivalence, but it is a proof gap or modeling error, not circularity: the conclusion is not assumed by the premises, and the resolution does not reduce to a self-citation or to a fitted input. Under the stated hard rules, the appropriate finding is no significant circularity.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

No new physical or model entities are postulated. The paper's quantitative claims instead rest on undisclosed cost coefficients, hand-chosen scenario parameters, and a structural reduction whose proof is sketched. 'Bounded flexibility' and work policies p1/p2/p3 are elements of the real hospital rule set, not invented entities.

free parameters (4)
  • Cost coefficients C_C (demand mismatch), C_R (request rejection), C_S (staffing), C_U (under-staffing), C_O… = not disclosed
    These weights define every objective value in Tables II and III; the claimed savings, VSS magnitudes, and the flexibility-regularity trade-off all scale with them, yet no numeric values are ever reported.
  • Reward temperature rho (Equation 9) = not disclosed
    Controls the trade-off between diversity and sampling quality in GNS; its value shapes the reported GNS gaps and is not given.
  • Scenario probabilities p_h = 0.6, p_l = 0.4 and demand shocks +4 and -2 = 0.6/0.4, +4/-2 per slot
    Chosen by hand in Section V C; they generate the scenario tree and therefore every VSS and GNS-versus-SP comparison. No sensitivity analysis is reported on these.
  • GNS training hyperparameters (architecture, learning rate, update frequency, top-k) = 50,000 epochs stated; rest not disclosed
    Required to reproduce the GNS results; only the epoch count is given in Algorithm 1.
assumptions (4)
  • domain assumption Block-separable recourse conditions of Birge and Louveaux apply to this model, making the multi-stage program reducible to a two-stage program without loss of accuracy
    Invoked in Theorem 1 and Remark 1 (Appendices I and J); the proof asserts rather than verifies the conditions against E7-E19, and the stage-0 work-policy indicators y^j_i (E9) couple all stages.
  • domain assumption Demand realizations are stage-wise independent and the scenario tree is fully known in advance
    Assumptions III B 2 and III B 3; the paper notes that data-driven scenario estimation is left to future work.
  • standard math GFN flow-proportional sampling theorem (Theorem 2)
    Quoted from Bengio et al. with the proof deferred to that reference (Appendix K).
  • domain assumption Demand coverage is evaluated against a fixed per-slot demand forecast with linear under- and over-supply penalties, plus an ad hoc skill-balance constraint
    Constraints C11 and E14 assume demand realizations and penalties are linear; the seniority-based skill coverage (Skill 1 to 5) was added for the real case in Section V E.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Nurse Staffing and Scheduling Problem with Bounded Flexibility and Demand Uncertainty." pith.science (2026). https://pith.science/paper/7UT7NXNL

@misc{pith2026250522124,
  author       = {Pith},
  title        = {Pith review of: A Nurse Staffing and Scheduling Problem with Bounded Flexibility and Demand Uncertainty},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7UT7NXNL}},
  note         = {Machine review of arXiv:2505.22124}
}
read the original abstract

Nurse staffing and scheduling are persistent challenges in healthcare due to demand fluctuations and individual nurse preferences. This study introduces the concept of bounded flexibility, balancing nurse satisfaction with strict rostering rules, particularly a real-world time regularity policy from a major hospital in Singapore. We model the problem as a multi-stage stochastic program to address evolving demand, optimizing both aggregate staffing and detailed scheduling decisions. A reformulation into a two-stage structure using block-separable recourse reduces computational burden without loss of accuracy. To solve the problem efficiently, we develop a Generative AI-guided algorithm. Numerical experiments with real hospital data show substantial cost savings and improved nurse flexibility with minimal compromise to schedule regularity. Numerical experiments based on real-world nurse profiles, nurse preferences, and patient demand data are conducted to evaluate the performance of the proposed methods. Our results demonstrate that the stochastic model achieves significant cost savings compared to the deterministic model. Notably, a slight reduction in the regularity level can remarkably enhance nurse flexibility.

Figures

Figures reproduced from arXiv: 2505.22124 by the authors.

Figure 1
Figure 1. FIG. 1. An illustrative example of work policy [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 4
Figure 4. FIG. 4. Actions sampling and schedule. For the current step, the action NurseB-Tuesday-P2 is drawn from the action space [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. FIG. 5. Shifts and time slots [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figures from the paper (2 more)
Figure 8
Figure 8. Figure 8: FIG. 8. Results of daily demand coverage gap by time slot. [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: FIG. 9. Results of daily demand coverage gap by skill. [PITH_FULL_IMAGE:figures/full_fig_p013_9.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

86 extracted references · 79 canonical work pages

  1. [1]

    Initially, a long-term staffing and scheduling plan is made, represented by the blue node in Figure 2, followed by a sequence of adjustments

    Decision Groups Within each stage h, there are two groups of decisions, namely the aggregate level decisions and detailed level decisions, separated by the realization of uncertainties (Figure 2). Initially, a long-term staffing and scheduling plan is made, represented by the blue node in Figure 2, followed by a sequence of adjustments. Within each stage ...

  2. [2]

    Now we introduce uncertainty realizations and scenar- ios of this problem

    Discrete Scenarios The uncertainty space can be captured by a finite set of discrete scenarios and the realizations of each stage are independent. Now we introduce uncertainty realizations and scenar- ios of this problem. A scenario tree is adopted to de- scribe the evolution process, assuming that the uncer- tainty space can be captured by a finite set o...

  3. [3]

    We assume the overall scenario tree can be estimated

    Scenario Tree The scenario tree is fully known for the entire planning horizon. We assume the overall scenario tree can be estimated. The scope of this paper is to capture the evolving uncer- tainties in the decision process. With the help of data- driven optimization, this study can be further improved with the gradually known information of uncertaintie...

  4. [4]

    Generative Flow Network (GFN) notations

  5. [5]

    Proposed Generative Nurse Scheduling (GNS) ap- proach

  6. [6]

    GFN is particularly good at sampling compositional, discrete objects like graphs or sets

    Generative flow networks (GFN) GFN is a novel generative modeling framework that trains models to sample diverse solutions proportional to the reward [10, 29]. GFN is particularly good at sampling compositional, discrete objects like graphs or sets. These objects are generated by a sequence of actions a, one part after the other, and ends at a terminal st...

  7. [7]

    First, the notation of consistent flow is intro- duced

    Training objective In this section, trajectory balance loss [30] is used as the training objective function for the above generative model. First, the notation of consistent flow is intro- duced. Then, the definition of trajectory balance loss is given. Assume that apart from the source node, a path τ visits several nodes s ∈ S(τ ). Let τ (s ′ ) denote a ...

    work page

  8. [8]

    This approach has two main features

    Generative nurse scheduling approach (GNS) Based on the concept of GFN discussed above, this section presents our generative nurse scheduling (GNS) approach in detail. This approach has two main features. First, utilizing a generative model, the nurse schedule is obtained via a sequence of actions, i.e., making an as- signment for a nurse. Each generated ...

    work page

Show all 86 references
  1. [9]

    The deterministic D1 solution is utilized as an initial solution and several ad- justments are made to it

    At each stage, with the hospital environment, nurse profile, average demand and deterministic model solution as the input information, we take Objective 9 as the re- ward function to train the model. The deterministic D1 solution is utilized as an initial solution and several ...

    2002

  2. [10]

    Halter, O

    M. Halter, O. Boiko, F. Pelone, C. Beighton, R. Har- ris, J. Gale, S. Gourlay, and V. Drennan, BMC health services research 17, 1 (2017)

    2017

  3. [11]

    Ringo, What healthcare cfos should know about the crna shortage (2021)

    2021

  4. [12]

    Breugem, T

    T. Breugem, T. Dollevoet, and D. Huisman, Management Science 68, 2619 (2022)

    2022

  5. [13]

    Ceschia, N

    S. Ceschia, N. Dang, P. De Causmaecker, S. Haspeslagh, and A. Schaerf, Annals of Operations Research 274, 171 (2019)

    2019

  6. [14]

    H. E. Miller, W. P. Pierskalla, and G. J. Rath, Operations Research 24, 857 (1976)

    1976

  7. [15]

    D. M. Warner, Operations Research 24, 842 (1976)

    1976

  8. [16]

    Aickelin and K

    U. Aickelin and K. A. Dowsland, Computers & opera- tions research 31, 761 (2004). 16

    2004

  9. [17]

    Kim and S

    K. Kim and S. Mehrotra, Operations Research 63, 1431 (2015)

    2015

  10. [18]

    J. R. Birge and F. Louveaux, Introduction to stochastic programming (Springer Science & Business Media, 2011)

    2011

  11. [19]

    Bengio, M

    E. Bengio, M. Jain, M. Korablyov, D. Precup, and Y. Bengio, Advances in Neural Information Processing Systems 34, 27381 (2021)

    2021

  12. [20]

    Bengio, S

    Y. Bengio, S. Lahlou, T. Deleu, E. J. Hu, M. Tiwari, and E. Bengio, Journal of Machine Learning Research 24, 1 (2021)

    2021

  13. [21]

    TTSH, (2024), accessed: 2024-08-10

    2024

  14. [22]

    S. Youn, H. N. Geismar, and M. Pinedo, Production and Operations Management 31, 4407 (2022)

    2022

  15. [23]

    Otero-Caicedo, C

    R. Otero-Caicedo, C. E. M. Casas, C. B. Jaimes, C. F. G. Garz´ on, E. A. Y. Vergel, and J. C. Z. Vald´ es, Operations Research for Health Care 38, 10.1016/j.orhc.2023.100389 (2023)

    2023

  16. [24]

    Schoenfelder, K

    J. Schoenfelder, K. M. Bretthauer, P. D. Wright, and E. Coe, European Journal of Operational Research 283, 390 (2020)

    2020

  17. [25]

    Legrain, J

    A. Legrain, J. Omer, and S. Rosat, European Journal of Operational Research 285, 196 (2020)

    2020

  18. [26]

    Maenhout and M

    B. Maenhout and M. Vanhoucke, Omega (United King- dom) 41, 903 (2013)

    2013

  19. [27]

    J. F. Bard and H. W. Purnomo, Socio-Economic Plan- ning Sciences 39, 193 (2005)

    2005

  20. [28]

    J. F. Bard and H. W. Purnomo, Journal of Scheduling 10, 5 (2007)

    2007

  21. [29]

    Guo and J

    J. Guo and J. F. Bard, Computers and Operations Re- search 138, 10.1016/j.cor.2021.105597 (2022)

    2022

  22. [30]

    Kheiri, A

    A. Kheiri, A. Gretsista, E. Keedwell, G. Lulli, M. G. Epitropakis, and E. K. Burke, Computers and Operations Research 130, 10.1016/j.cor.2021.105221 (2021)

    2021

  23. [31]

    Z. Chen, Y. Dou, and P. D. Causmaecker, Computers and Industrial Engineering 171, 10.1016/j.cie.2022.108430 (2022)

    2022

  24. [32]

    F. F. Oberweger, G. R. Raidl, E. R¨ onnberg, and M. Hu- ber, Artificial Intelligence, and Operations Research 2022, 300 (2022)

    2022

  25. [33]

    J. Park, J. Chun, S. H. Kim, Y. Kim, and J. Park, Inter- national Journal of Production Research 59, 3360 (2021)

    2021

  26. [34]

    D. W. Zhang, C. Rainone, M. Peschl, and R. Bondesan, arXiv preprint arXiv:2302.05446 (2023)

    2023 arXiv

  27. [35]

    Zhang, W

    X. Zhang, W. Zhao, Y. Sun, J. Sun, Y. Shen, X. Dong, and Z. Yang, in ISSTA 2023 - Proceedings of the 32nd ACM SIGSOFT International Symposium on Software Testing and Analysis (Association for Computing Ma- chinery, Inc, 2023) pp. 942–953

    2023

  28. [36]

    S. Liu, Q. Cai, Z. He, B. Sun, J. McAuley, D. Zheng, P. Jiang, and K. Gai, in Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (Association for Computing Machinery, 2023) pp. 1524–1534

    2023

  29. [37]

    M. Chan, Y. Yin, B. Amado, P. Williams, and D. Xiao, Cornell University Computational Optimization Open Textbook (2020)

    2020

  30. [38]

    Bengio, S

    Y. Bengio, S. Lahlou, T. Deleu, E. J. Hu, M. Tiwari, and E. Bengio, Journal of Machine Learning Research 24, 1 (2023)

    2023

  31. [39]

    Malkin, M

    N. Malkin, M. Jain, E. Bengio, C. Sun, and Y. Bengio, Advances in Neural Information Processing Systems 35, 5955 (2022)

    2022

  32. [40]

    Wikipedia, (2024), accessed: 2024-08-10

    2024

  33. [41]

    Anderson, M

    M. Anderson, M. Bodur, S. Rathwell, and V. Sarhangian, INFORMS Journal on Applied Analytics 53, 133 (2023)

    2023

  34. [42]

    C. W. Chan, M. Huang, and V. Sarhangian, Operations Research 69, 1936 (2021)

    2021

  35. [43]

    E. B. B¨ oovarsd´ ottir, P. Smet, G. V. Berghe, and T. J. Stidsen, European Journal of Operational Research 292, 980 (2021)

    2021

  36. [44]

    D. S. Lai, J. M. Leung, W. Dullaert, and I. Marques, Eu- ropean Journal of Operational Research 284, 285 (2020)

    2020

  37. [45]

    Strandmark, Y

    P. Strandmark, Y. Qu, and T. Curtois, Computers and Operations Research 120, 10.1016/j.cor.2020.104945 (2020)

    2020

  38. [46]

    Wolbeck, N

    L. Wolbeck, N. Kliewer, and I. Marques, European Jour- nal of Operational Research 284, 1121 (2020)

    2020

  39. [47]

    T. I. Wickert, P. Smet, and G. V. Berghe, Computers and Operations Research 104, 319 (2019)

    2019

  40. [48]

    T. C. Wong, M. Xu, and K. S. Chin, Computers and Operations Research 51, 99 (2014)

    2014

  41. [49]

    Maenhout and M

    B. Maenhout and M. Vanhoucke, Omega (United King- dom) 41, 485 (2013)

    2013

  42. [50]

    Martin, D

    S. Martin, D. Ouelhadj, P. Smet, G. V. Berghe, and E. ¨Ozcan, Expert Systems with Applications 40, 6674 (2013)

    2013

  43. [51]

    P. D. Wright and S. Mahar, Omega (United Kingdom) 41, 1042 (2013)

    2013

  44. [52]

    Vanhoucke and B

    M. Vanhoucke and B. Maenhout, European Journal of Operational Research 196, 457 (2009)

    2009

  45. [53]

    Nishikawa-Toomey, T

    M. Nishikawa-Toomey, T. Deleu, J. Subramanian, Y. Bengio, and L. Charlin, arXiv (2022)

    2022

  46. [54]

    M. Jain, S. C. Raparthy, A. Hern´ andez-Garcıa, J. Rector- Brooks, Y. Bengio, S. Miret, and E. Bengio, in Interna- tional Conference on Machine Learning (PMLR, 2023) pp. 14631–14653

    2023

  47. [55]

    Lahlou, T

    S. Lahlou, T. Deleu, P. Lemos, D. Zhang, A. Volokhova, A. Hern´ andez-Garcıa, L. N. Ezzine, Y. Bengio, and N. Malkin, in International Conference on Machine Learning (PMLR, 2023) pp. 18269–18300

    2023

  48. [56]

    L. Pan, D. Zhang, M. Jain, L. Huang, and Y. Bengio, arXiv preprint arXiv:2302.09465 (2023)

    2023 arXiv

  49. [57]

    Z. A. Abdalkareem, A. Amir, M. A. Al-Betar, P. Ekhan, and A. I. Hammouri, Health and Technology 11, 445 (2021)

    2021

  50. [58]

    Sitompul and S

    D. Sitompul and S. U. Randhawa, Journal of the Society for Health Systems 2, 62 (1990)

    1990

  51. [59]

    Jaumard, F

    B. Jaumard, F. Semet, and T. Vovor, European journal of operational research 107, 1 (1998)

    1998

  52. [60]

    M. A. H. Dempster, M. L. Fisher, L. Jansen, B. J. Lageweg, J. K. Lenstra, and A. H. Rinnooy Kan, Math- ematics of Operations Research 8, 525 (1983)

    1983

  53. [61]

    A. J. Kleywegt, A. Shapiro, and T. Homem-de Mello, SIAM Journal on optimization 12, 479 (2002)

    2002

  54. [62]

    Flinkman, H

    M. Flinkman, H. Leino-Kilpi, and S. Salanter¨ a, Journal of advanced nursing 66, 1422 (2010)

    2010

  55. [63]

    Zaerpour, M

    F. Zaerpour, M. Bijvank, H. Ouyang, and Z. Sun, Pro- duction and Operations Management 31, 645 (2022)

    2022

  56. [64]

    Rath and K

    S. Rath and K. Rajaram, Production and Operations Management 31, 1271 (2022)

    2022

  57. [65]

    K. Sun, M. Sun, D. Agrawal, R. Dravenstott, F. Rosinia, and A. Roy, Production and Operations Management32, 3699 (2023). 17 Appendix A: Abbreviations TABLE IV. Abbreviations Abbreviation Definition AM morning slot PM afternoon slot N night slot NW nonworking day W working day ...

    2023

  58. [66]

    deterministic ✓ ✓ ✓ MOP NSGAII and policy-based heuristics

  59. [67]

    deterministic ✓ HOM hybrid exact and heuristic method

  60. [68]

    stochastic ✓ ✓ ✓ MIP ϵ-constraint solution method

  61. [69]

    stochastic ✓ ✓ two-stage SP L-shaped method

  62. [70]

    stochastic ✓ ✓ two-stage SP SAA-based heuristic

  63. [71]

    deterministic ✓ MIP column generation-based heuristic

  64. [72]

    deterministic ✓ ✓ NN heuristic strategy

  65. [73]

    deterministic ✓ ✓ ✓ MOP lexicographic goal programming with acceptance threshold

  66. [74]

    weekly demand ✓ hidden MDP local search and perturbation heuristics

  67. [75]

    stochastic ✓ ✓ multi-stage SP

  68. [76]

    deterministic ✓ NFM graph construction algorithm and solver

  69. [77]

    weekly demand ✓ ✓ DP online stochastic algorithm

  70. [78]

    deterministic ✓ IP column generation-based heuristic

  71. [79]

    deterministic ✓ IP commercial solver

  72. [80]

    nurse absence IP variable neighborhood descent heuristic

  73. [81]

    stochastic ✓ two-stage SP integer L-shaped algorithm

  74. [82]

    deterministic ✓ two-stage model hybrid heuristic

  75. [83]

    deterministic ✓ ✓ IP branch-and-price

  76. [84]

    deterministic ✓ ✓ IP branch-and-price and meta- heuristic

  77. [85]

    deterministic ✓ bicriteria IP commercial solver

  78. [86]

    deterministic ✓ agent-based model meta-heuristic

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.