v0.3.0

SDDP v0.3.0

Diff since v0.2.4

Merged pull requests:

• Updates for JuMP 0.21 (#300) (@odow)
• Install TagBot as a GitHub Action (#303) (@JuliaTagBot)

v0.2.4

v0.2.4 (2020-02-06)

Diff since v0.2.3

Merged pull requests:

• Add a counter for total solves (#301) (odow)
• Add tests to improve code coverage (#299) (odow)
• Update formatter (#298) (odow)

v0.2.3

v0.2.3 (2020-01-23)

Diff since v0.2.2

Merged pull requests:

• Fix deterministic for affine expr with no terms (#297) (odow)
• Move test for primal feasibility earlier (#296) (odow)
• Allow offsets in risk measures (#294) (odow)
• Add example for IJOC paper (#293) (odow)

v0.2.2

v0.2.2 (2020-01-09)

Diff since v0.2.1

Closed issues:

• Pkg> test SDDP fails (#288)

Merged pull requests:

• Fix flaky time limit in tests (#291) (odow)
• Update copyright (#290) (odow)
• Remove MathOptFormat as a dependency (#289) (odow)

v0.2.1

v0.2.1 (2019-12-18)

Diff since v0.2.0

Closed issues:

• Anonymous JuMP Variables (#250)
• File writing for non-MILPs (#239)
• Add tools to visualize value function (#223)
• Initial Markov State for every variable (#211)
• Scenario generation (#208)
• Write a beginners SDDP for dummies tutorial (#65)

Merged pull requests:

• Test on Julia 1.3 (#287) (odow)
• Tools to visualize the value function: Part II (#286) (odow)
• Add catch when budget < # stages (#285) (odow)
• Write MOF files by default (#284) (odow)
• Improve docs for MarkovianPolicyGraph (#283) (odow)
• Add support for approximating Markov lattice (#282) (odow)
• Add documentation for asynchronous mode (#281) (odow)
• Bump minimum HTTP version (#278) (odow)
• WIP: tools to visualize the value function (#272) (odow)

v0.2.0

v0.2.0 (2019-12-14)

Diff since v0.1.0

Closed issues:

• Setting lower_bound when sense is :Max (#270)
• Modifying existing cut coefficients (#267)
• Add docs on simulating policy graphs with dummy nodes (#266)
• Timeout in dashboard (#226)
• SLDP Ideas (#210)
• Proposal: parallelisation schemes (#196)
• Kokako to-do list (#181)

Merged pull requests:

• Add documentation on simulating non-standard policy graphs (#280) (odow)
• Add rollout feature (#279) (odow)
• Bump minimum HTTP version (#278) (odow)
• WIP: asynchronous mode (#277) (odow)
• JuliaFormatter updates (#276) (odow)
• Add upper bounds to Project.toml (#274) (odow)
• Add JuliaFormatter to the codebase (#273) (odow)
• Provide better error messages in PolicyGraph when providing bounds (#271) (odow)

v0.1.0

v0.1.0 (2019-10-10)

Diff since v0.0.2

Closed issues:

• Out-of-sample simulation with objective states (#251)
• Julia 1.1 Issue (#249)
• Recorded Simulate Variables (#243)
• Expose machine-readable log (#235)
• BoundsError: attempt to access 0-element Array{SDDP.Log,1} at index [0] (#231)
• Statistical stopping rule (#204)
• rhsnoise to SDDP.parametrize (#200)
• Provide Lipschitz values for ambiguity sets in constructor (#199)
• SDDiP (#195)
• SDDP.write_subproblem_to_file (#193)
• Add cut_deletion_minimum argument to train (#188)
• Add de-bugging docs (#187)
• Saving trained model (#184)
• SDDP.simulate with vectors (#183)
• Installing SDDP (#182)
• Noise selection in backward pass to reduce amount of subproblems (#177)
• 2-dim tuple for noise (#176)
• About dual multipliers of a constraint (#175)
• infeasible solution (#174)
• SDDP.jl, JuMP 0.19, and Julia 1.0 (#172)
• Documentation for publication plots (#168)
• Style Fixes (#166)
• Fix Wasserstein Properly (#162)
• Some comments about the package (#158)
• Get dual variables out of the simulation (#156)
• Add infeasibility cuts (#151)
• Accessing the stage-objective in a simulation (#150)
• Document overloadable functions (#149)
• Constraint depending on state value (#126)
• Proposal: Stopping Rules (#124)
• Refactor (#118)
• Add a check for numerical stability (#115)
• Proposal: change noise syntax (#107)
• Non-convex controls (#100)
• State dependent lipschitz constant (#96)
• Proposal: More Exotic Policy Graphs (#92)
• Add test for async cut writing (#91)
• Release proper tagged versions (#85)
• Rename ':noise' key in solution storage (#84)
• Proposal: Sampling oracles (#80)
• Best way to use historical inflow data in dams model (#79)
• Stagewise indepdendent noise in constraint matrix (#70)
• Tidy up d3 visualiser (#36)
• Write docs in Documenter.jl (#30)
• Add ability to sample forward pass with different probabilities (#29)
• Remove the need to specify objective\_bound (#28)
• A better way to run simulations using different noises to the original model (#10)
• Numerical Issues (#6)

Merged pull requests:

• Update Project.toml and using's for registration (#263) (odow)
• Add explanation of do syntax in docs (#262) (odow)
• Update hydro tutorial (#260) (odow)
• Performance and style improvements (#259) (odow)
• Use in-place operations to improve performance (#258) (odow)
• Fix headers in docs (#255) (odow)
• Refactor documentation into tutorials and how-to guides (#254) (odow)
• Update docs for JuMP 0.20 (#253) (odow)
• Implement out-of-sample sampling scheme (#252) (odow)
• Update add (#248) (Thuener)
• Add function to save machine-readable log (#247) (odow)
• Fix custom_recorders example (#245) (Thuener)
• Allow skip_undefined_variables in simulate (#244) (odow)
• Updated hydro valley example and tests (#240) (conrad-mac)
• WIP: SDDiP (#237) (lkapelevich)
• Enforce [0,1] bounds on binary variables for LP relaxation (#236) (bfpc)
• Implement work-around for bug in RecipesBase (#234) (odow)
• Remove Manifest.toml now MOF is registered (#233) (odow)
• Fix a number of issues with too few iterations in the log (#232) (odow)
• Fix failing doctests (#230) (odow)
• Add beginnings of a deterministic equivalent solver (#229) (odow)
• Add ORL paper example (#228) (odow)
• Dashboard warning (#227) (odow)
• Add a visualization dashboard (#225) (odow)
• Rp/backward (#224) (roporte)
• [bug] fix bug in modify_similar_nodes (#221) (odow)
• Tidy the documentation and signature of train (#220) (odow)
• Add a settable optimize_hook with access to the current scenario. (#219) (odow)
• Fix absolute value representation and update bound (#218) (bfpc)
• Add the tiger problem as an example (#217) (odow)
• Simplify the belief backward pass (#215) (odow)
• Update SLDP examples with actual random values (#214) (odow)
• Add examples from Ahmed's Stochastic Lipschitz Dynamic Programming paper (#213) (odow)
• Add an infinite-horizon example (#212) (odow)
• Add MSPPy hydro-thermal example (#209) (odow)
• Add more binary examples (#207) (odow)
• Enable multistage mixed integer problems via conv relaxation (#206) (odow)
• Modified Statistical stopping rule. (#205) (aferragu)
• Refactor passing of lipschitz values for belief (#203) (odow)
• Add more examples of (at)rhsnoise constraints to upgrading guide (#202) (odow)
• Update docs by adding color (#201) (odow)
• Hash risk-set cuts in multi-cut (#197) (odow)
• Add extension dictionaries for extensions (#194) (odow)
• Add belief states (#192) (odow)
• Add farmer example using Literate.jl (#191) (odow)
• Add cut_deletion_minimum to train (#190) (odow)
• Add a debugging tutorial (#189) (odow)
• Improve code coverage (#186) (odow)
• Implement multi-cut (#185) (odow)
• Merge kokako (#180) (odow)
• Make some changes to fix using Gurobi asynchronously with constraint … (#173) (odow)
• Modify example for ORL revision (#171) (odow)
• Add more docs on publicationplot (#169) (odow)
• Update .travis.yml (#167) (odow)
• Update style of risk measure code. (#165) (odow)
• Cut oracle style fixes (#164) (odow)

Pin for MPC

Updates to travis (#163)

* Upper bound Julia and JuMP prior to rewrite

* Simplify travis

* I forgot about needing gfortran for Ipopt

v0.0.1

Tidy up examples and reduce amount of printing in tests (#119)