What's New in MOST 1.3.1

Released Jul 12, 2025

Below is a summary of the changes since version 1.3 of MOST. See the CHANGES.md file for all the gory details. For release notes for previous versions, see Appendix B of the MOST User's Manual.

Changes:

• most_summary() now skips display of non-existent contingencies.
• Include HiGHS solver (via HiGHSMEX interface), if available, in unit commitment tests.

Bugs Fixed:

• Fix issue #45 where most() does not properly handle cases with contingencies defined only in some periods/scenarios. Thanks to Stefano Nicolin.
• Fix issue with most_summary() when ramp results are missing.
• Tweak tests to work around bug in HiGHS-based linprog and intlinprog in Optimization Toolbox R2024a and R2024b.
• Fix issue caused by tiny non-zero values for commitment variables. Don't count on MP-Opt-Model's miqps_<solver>() functions to round integer variable solutions.

PRO Version

• There is now a PRO version of MOST that adds support for MATPOWER DC lines as described in Section 7.6.3 of the MATPOWER User's Manual. Please contact [email protected] for information on obtaining this version.

What's New in MOST 1.3

Released May 10, 2024

Below is a summary of the changes since version 1.2 of MOST. See the
CHANGES.md file for all the gory details. For release notes for
previous versions, see Appendix B of the MOST User's Manual.

New Features:

• New Sphinx-based Reference documentation.

Changes:

• Reduce memory requirements for long horizon cases with storage by
  forming/storing transposes of matrices for storage constraints.
  Requires MP-Opt-Model 4.2 or later.
• Speed up building unit commitment (min up/down time) constraints.
  Improvement can be quite substantial on large problems.

Bugs Fixed:

• Fix issue #37 which caused a fatal error in storage input
  checks with multiple storage units under some circumstances.
  Thanks to Keir Steegstra.
• Fix issue #39 in which the value of mdi.Delta_T, the number
  of hours represented by each period, was not being accounted for in
  most of the terms in the objective function.
  Thanks to Stefano Nicolin.

Incompatible Changes:

• Remove extra column in mdo.results.ExpectedRampCost and ignore for
  single period.

What's New in MOST 1.2

Released Dec 13, 2022

Below is a summary of the changes since version 1.1 of MOST. See the
CHANGES.md file for all the gory details. For release notes for
previous versions, see Appendix B of the MOST User's Manual.

Changes:

• Ramping reserves and constraints are now included for the
  transition from the initial state into period 1, except for
  single-period problems.
• Added calculation of expected TLMP (temporal locational marginal
  price) based on work by Guo, Chen, Tong in [1,
  2, 3]. For generators, these are
  returned in mdo.results.GenTLMP and mdo.results.CondGenTLMP.
  For storage units they are returned in mdo.results.StorageTLMPc,
  mdo.results.StorageTLMPd, mdo.results.CondStorageTLMPc, and
  mdo.results.CondStorageTLMPd. See Table 5-13 in the
  MOST User's Manual.
• For deterministic cases with storage where ForceCyclicStorage is 0,
  ensure that initial storage bounds are equal to initial storage and
  output a warning if they are modified. Fix deterministic UC tests
  where this was causing results to change depending on value of rho.

Bugs Fixed:

• Plotting of commitment schedule using plot_uc() did not work
  properly in a subplot, such as in t_most_uc(). Thanks to Lim Han.
• Fix tests that were failing under Octave 7.x.
• Fix issue #29 where a typo caused a check on md.UC.MinDown > 1
  to be skipped. Thanks to Talha Iqbal.

Incompatible Changes:

• Modified definition of ramping reserves for period t (and all
  corresponding input and output parameters) to refer to the transition
  from t-1 to t, not t to t+1. This means that the ramping
  reserves for the transition into the first period are now optimization
  variables and the corresponding constraints are explicit. This is for
  multiperiod problems only. Ramping reserves and contraints are explicitly
  excluded for single-period problems.
  Note: This change also corrects an error in (4.11), where \gamma^t
  is now correct. Previously it should have been \gamma^{t+1}, as it was
  in the code.

References

1. Y. Guo, C. Chen and L. Tong, “Pricing Multi-Interval Dispatch Under
   Uncertainty Part I: Dispatch-Following Incentives,” IEEE Transactions
   on Power Systems, vol. 36, no. 5, pp. 3865–3877, Sept. 2021,
   doi: 10.1109/TPWRS.2021.3055730.

2. C. Chen, Y. Guo and L. Tong, “Pricing Multi-Interval Dispatch Under
   Uncertainty Part II: Generalization and Performance,” IEEE Transactions
   on Power Systems, vol. 36, no. 5, pp. 3878–3886, Sept. 2021,
   doi: 10.1109/TPWRS.2020.3045162.

3. C. Chen and L. Tong, “Pricing Real-time Stochastic Storage Operations,”
   2022 Power Systems Computation Conference (PSCC), June 27–July 1, 2022,
   doi: 10.48550/arXiv.2204.08140.

What's New in MOST 1.1

Released Oct 8, 2020

Below is a summary of the changes since version 1.0.2 of MOST. See the
CHANGES.md file for all the gory details. For release notes for
previous versions, see Appendix B of the MOST User's Manual.

Changes:

• Requires MATPOWER 7.1 or later.
• Output of most_summary() includes sections for fixed loads and for
  expected stored energy for storage units.
• Relies on MP-Opt-Model 3.0, which can be found at
  https://github.com/MATPOWER/mp-opt-model and is included in
  MATPOWER 7.1.
  • Significant performance improvement for some problems when constructing
    sparse matrices for linear constraints or quadratic costs (e.g. during
    problem setup).
    Thanks to Daniel Muldrew.
  • Uses the @opt_model/solve() method rather than calling
    miqps_matpower() or qps_matpower() directly.
  • Uses the @opt_model/get_soln() method to extract variable and shadow
    price results, rather than doing the indexing manually.

Bugs Fixed:

• Fix bug #6 where building a model without solving it, or solving a
  previously built model resulted in a fatal error.
  Thanks to Baraa Mohandes.
• Fix bug #11 where storage constraints were not correct for
  t=1 and rho not equal to 1. Thanks to Baraa Mohandes.
• Fix issue #16, where the om field of the output MOST data struct
  (mdo) was a handle to the same object as as the om field of the
  input MOST data struct (mdi), meaning that changing one would modify
  the other. Thanks to Baraa Mohandes.

Incompatible Changes:

• Objective function value returned in mdo.QP.f updated to include the
  previously missing constant term.

What's New in MOST 1.0.2

Released Jun 20, 2019

Below is a summary of the changes since version 1.0.1 of MOST. See the
CHANGES.md file for all the gory details. For release notes for
previous versions, see Appendix B of the MOST User's Manual.

Bugs Fixed:

• Fix default solver selection issue in t_most_w_ds test.

Other Changes:

• Add CITATION file.
• Other miscellaneous documentation updates, e.g. MATPOWER website
  links updated to https://matpower.org, separate references for
  MATPOWER software and User's Manual, with DOIs.

What's New in MOST 1.0.1

Released Oct 30, 2018

Below is a summary of the changes since version 1.0 of MOST. See the
CHANGES.md file for all the gory details. For release notes for
previous versions, see Appendix B of the MOST User's Manual.

Bugs Fixed:

• Fix bugs in plot_uc_data() resulting in incorrect legends.
• Fix dimension of RampWear cost indexing if mdi.OpenEnded is true.
• Add missing constant term to objective function value reported by
  most_summary.

Other Changes:

• LaTeX source code for MOST User's Manual included in docs/src.
• Updated to use OOP notation for opt_model object, and avoid calls
  to deprecated methods, using init_indexed_name() and
  add_lin_constraint() instead.
• Updated to use MATPOWER's new quadratic costs in opt_model in
  place of the legacy cost model.

Incompatible Changes:

• Failure of the optimization no longer halts execution and jumps to
  the debugger.
• Requires MATPOWER 7.x or later.