v0.4.5

SDDP v0.4.5

Diff since v0.4.4

Closed issues:

• Train model with Asynchronous mode (#492)
• Cut-sharing property (#493)
• SDDP.simulate with known realization (#494)
• set_silent does not work in a subproblem (#495)
• Error after updating to SDDP.jl 0.4.4 (#497)
• Small typo (#498)
• Simulating the policy section (#499)
• Objective states - error in the example (#503)
• GARCH? (#504)
• Is there a way to efficiently add deterministic node in policy graph? (#508)
• Constraint information log in SDDP (#509)
• Update to JuMP v0.23 (#512)

Merged pull requests:

• Fix #498 (#500) (@odow)
• Fix typos in First Steps (#501) (@odow)
• Fix test failures (#502) (@odow)
• [docs] fix typo in objective state tutorial (#506) (@odow)
• [docs] add auto-regressive tutorial (#507) (@odow)
• Fix issue with set_silent in a subproblem (#510) (@odow)
• Typo \cup-\cap (#511) (@bfpc)
• Update to JuMP v0.23 (#514) (@odow)

v0.4.4

SDDP v0.4.4

Diff since v0.4.3

Closed issues:

• Add cut selection for objective and belief states (#326)
• read_cuts_from_file does not load cuts into the BellmanFunction (#360)
• Add code from biobjective paper (#461)
• Loading and writing cuts from files (#467)
• Historical Simulation (#485)

Merged pull requests:

• WIP: entropic paper (#460) (@odow)
• WIP: biobjective paper (#463) (@odow)
• Save states in write_cuts_to_file (#468) (@odow)
• Fix BoundStalling for deterministic policy (#470) (@odow)
• Add BanditDuality handler (#471) (@odow)
• Fixed typos (#472) (@vfdev-5)
• Minor formatting updates (#473) (@odow)
• Fix behavior of BoundStalling stopping rule (#474) (@odow)
• Add scripts for A/B benchmarking (#475) (@odow)
• Fix the benchmark script and add a model (#476) (@odow)
• Switch to a functional form for the tests (#478) (@odow)
• Refactoring of bellman functions for concave cut selection (#479) (@odow)
• Update brazilian example (#480) (@odow)
• Add extra cut to Lagrangian problem (#481) (@odow)
• Another improvement for lagrangian (#482) (@odow)
• Fix magnitude warning with zero coefficients (#483) (@odow)
• Remove bound-hack in StrengthenedConicDuality (#486) (@odow)
• Use modified BFGS to solve Lagrangian (#487) (@odow)
• BuMP to JuMP v0.22 (#489) (@odow)
• Update benchmarks (#490) (@odow)

v0.4.3

SDDP v0.4.3

Diff since v0.4.2

Merged pull requests:

• Fix typo in risk tutorial (#459) (@odow)
• Add a biobjective solver (#462) (@odow)
• Re-organize how paper materials are stored (#464) (@odow)
• Move remaining examples (#465) (@odow)
• Add forward_pass_callback (#466) (@odow)

v0.4.2

SDDP v0.4.2

Diff since v0.4.1

Merged pull requests:

• Fix bug in lagrangian duality (#457) (@odow)

v0.4.1

SDDP v0.4.1

Diff since v0.4.0

Merged pull requests:

• Use tigher bounds for the incoming states (#454) (@odow)
• Add a second solve to find a sparse dual vector (#455) (@odow)
• Update Project.toml (#456) (@odow)

v0.4.0

SDDP v0.4.0

Diff since v0.3.17

Closed issues:

• Decision-Hazard variables (#238)
• Add a callback to modify data over the iterations (#416)
• Use integer forward pass (#431)
• Documentation feedback (#432)
• Document how to access individual subproblems (#435)
• Add better documentation on relatively complete recourse (#436)
• Historical Simulation - BoundsError (#440)
• Integer variables when constructing a PolicyGraph (#441)
• Fix $ escaping in documentation (#442)
• Objective state with state-dependent Ω (#444)
• Fix docs of historical sampling scheme (#446)
• Installation issue (#451)

Merged pull requests:

• Fix dollar escaping in docs (#447) (@odow)
• [docs] add relatively complete recourse section (#448) (@odow)
• [breaking] Large refactoring of stochastic integer programs (#449) (@odow)
• [docs] fix documentation of Historical sampling scheme (#450) (@odow)
• Rename ConicDuality and relax_integrality (#453) (@odow)

v0.3.17

SDDP v0.3.17

Diff since v0.3.16

Closed issues:

• Forward sample with risk-adjusted probabilities (#323)
• Tests are flakey (#427)
• Add more problem descriptors (#430)

Merged pull requests:

• Add PSRSamplingScheme (#426) (@odow)
• Various small improvements to the documentation (#433) (@odow)
• Restructure the tutorials (#437) (@odow)
• Display more attributes of a model (#438) (@odow)
• Rewrite intro tutorial to add uncertainty (#439) (@odow)

v0.3.16

SDDP v0.3.16

Diff since v0.3.15

Closed issues:

• Add a warning if you try to train a model with existing cuts (#305)
• Historical ForwardPass sampling scheme (#414)
• Function to add back all cuts (#415)
• Add example using Gurobi in parallel (#417)

Merged pull requests:

• Add RiskAdjustedForwardPass (#413) (@odow)
• Update risk measure docstrings (#418) (@odow)
• Allow Historical to sample sequentially (#420) (@odow)
• Add warning when training with existing cuts (#421) (@odow)
• Add add_all_cuts (#422) (@odow)
• Update improve_computational_performance.md (#423) (@odow)
• Update docs to build on 1.6 (#425) (@odow)

v0.3.15

SDDP v0.3.15

Diff since v0.3.14

Closed issues:

• Check that the initial root state respects variable bounds (#387)
• Don't bound stall in initial iterations if no progress and Simulation is != Bound (#394)
• Unable to access water values in multi storage SDDP calculation (#398)
• MathOptFormat for SDDP.jl (#399)
• Sanitize stageobjective macro (#404)

Merged pull requests:

• Add docs on accessing previous stage (#406) (@odow)
• Sanitize inputs to stageobjective macro (#407) (@odow)
• Expand docs for water values (#408) (@odow)
• Silent by default (#409) (@odow)
• Add StoppingChain (#410) (@odow)
• Check feasibility of initial point (#411) (@odow)
• Add JuliaFormatter (#412) (@odow)

v0.3.14

SDDP v0.3.14

Diff since v0.3.13

Closed issues:

• Add guidance to documentation to solve large model (#392)

Merged pull requests:

• Fix O(n^2) function in algorithm.jl (#393) (@odow)