POMDPs.jl Versions Save

MDPs and POMDPs in Julia - An interface for defining, solving, and simulating fully and partially observable Markov decision processes on discrete and continuous spaces.

v0.9.6

6 months ago

POMDPs v0.9.6

Diff since v0.9.5

Merged pull requests:

  • CompatHelper: bump compat for NamedTupleTools to 0.14 for package POMDPTools, (keep existing compat) (#427) (@github-actions[bot])
  • CompatHelper: bump compat for UnicodePlots to 3 for package POMDPTools, (keep existing compat) (#428) (@github-actions[bot])
  • Fix for underlying MDP not preserving state-dependent action space (#430) (@AlexBork)
  • Emphasize state immutability in the warning. (#431) (@BoZenKhaa)
  • Fix a minor typo in docstring. (#438) (@aidmandorky)
  • Remove abstract types from rollout simulator (#441) (@BoZenKhaa)
  • add initialstate for SparseTabular(PO)MDP (#445) (@fredcallaway)
  • Add state-dep-actions support for EpsGreedyPolicy (#450) (@NeroBlackstone)
  • state-dep-action document improvement (#451) (@NeroBlackstone)
  • Add ParticleFilterTrees and NativeSARSOP to README (#457) (@WhiffleFish)
  • EpsGreedyPolicy documentation update (#458) (@NeroBlackstone)
  • Changing from Travis to CI (#461) (@felixmg312)
  • Updated CI's for all libraries except VDPTag in Readme.md (#462) (@felixmg312)
  • Remove Deprecations (#466) (@WhiffleFish)
  • Update CI.yml (#479) (@zsunberg)
  • CompatHelper: bump compat for StatsBase to 0.34 for package POMDPTools, (keep existing compat) (#483) (@github-actions[bot])
  • FullyObservablePOMDP obsindex bugfix (#485) (@WhiffleFish)
  • Bring MacOS/Windows back into CI (#486) (@WhiffleFish)
  • Update README.md (#487) (@zsunberg)
  • Dict policy (#491) (@NeroBlackstone)
  • Playback default (#493) (@zsunberg)
  • Update def_updater.md (#494) (@dykim07)
  • added atol to all distribution checks (#498) (@zsunberg)
  • removes deprecated CommonRLInterface.@provide (#499) (@zsunberg)
  • Link to tiger paper (#500) (@zsunberg)
  • cleaned up test tests (#501) (@zsunberg)
  • use CommonRLInterface.AutomaticDefault when possible (#502) (@zsunberg)
  • Use Random.default_rng() (#503) (@NeroBlackstone)
  • Update imports in README.md example (#506) (@BoZenKhaa)
  • Update README.md (#508) (@zsunberg)
  • Fix Z vs. O typo (#512) (@lassepe)
  • Remove latest observation from CommonRLEnv POMDP state (#516) (@johannes-fischer)
  • Add generative interface to FullyObservablePOMDP (#517) (@johannes-fischer)
  • Replace @kwdef with Base.@kwdef to make the package compatible with older Julia versions (#518) (@himanshugupta1009)
  • Update README.md (#520) (@BoZenKhaa)
  • Fix pomdpmodels 97 (#522) (@zsunberg)
  • Bump version (#523) (@zsunberg)

Closed issues:

  • Consolidate Packages into POMDPTools.jl (#290)
  • Add SparsePFT (#384)
  • Upgrade ParticleFilters to use POMDPTools (#398)
  • Upgrade LocalApproximationValueIteration to use POMDPTools (#399)
  • Upgrade QMDP to use POMDPTools (#402)
  • Upgrade FIB to use POMDPTools (#403)
  • Upgrade ARDESPOT to use POMDPTools (#407)
  • Upgrade POMCPOW to use POMDPTools (#412)
  • CompatHelper does not work on POMDPTools (#426)
  • State-dependent action spaces in POMDP are not transfered to underlying MDP (#429)
  • Any example about implicit definition of state space and action space? (#433)
  • 'POMDPTools' is not installable via general registry (#437)
  • belief-dependent transition function (#439)
  • Exporting alpha vectors. (#442)
  • SparseTabularMDP doesn't support initialstate (#444)
  • Add two function to EpsGreedyPolicy let it select action by current state. (#449)
  • Function convert_s error (#452)
  • @POMDP_require deprecation in POMDPTools (#463)
  • Build is broken (#469)
  • Remove type parameters for MDP and POMDP (#475)
  • CI badges look scary (#478)
  • FullyObservablePOMDP broken obsindex (#480)
  • Not actually testing on MacOS/Windows (#481)
  • Use Random.default_rng() instead of Random.GLOBAL_RNG (#496)
  • Typo in documentation (#511)
  • Docs not building (#519)

v0.9.5

1 year ago

POMDPs v0.9.5

Diff since v0.9.4

Closed issues:

  • Model update checklist (#313)
  • Add AdaOPS (#383)
  • Any recommendation to fix this? (#388)
  • transition and observation return distributions, not states or observations (#390)
  • PRs for tools to point to POMDPTools (#392)
  • Accept PRs for tools to point to POMDPTools (#393)
  • Upgrade X to use POMDPTools (#395)
  • Upgrade X to use POMDPTools (#396)
  • Upgrade DiscreteValueIteration to use POMDPTools (#397)
  • Upgrade MCTS to use POMDPTools (#401)
  • Upgrade SARSOP to use POMDPTools (#405)
  • Upgrade BasicPOMCP to use POMDPTools (#406)
  • Upgrade PointBasedValueIteration to use POMDPTools (#414)
  • Upgrade TabularTDLearning to use POMDPTools (#415)
  • Upgrade DeepQLearning to use POMDPTools (#416)
  • Update POMDPs.jl README to talk about POMDPTools (#418)
  • Update POMDPs.jl documentation front matter to talk about POMDPTools (#419)
  • Fix POMDPs.jl CI (#420)
  • Transition from LightGraphs.jl to Graphs.jl (#424)

Merged pull requests:

  • add AdaOPS to POMDP solvers in README (#386) (@WhiffleFish)
  • Pomdptools (#391) (@zsunberg)
  • fix #424 (#425) (@zsunberg)

v0.9.4+docs

1 year ago

v0.9.4

2 years ago

POMDPs v0.9.4

Diff since v0.9.3

Closed issues:

  • Checklist for Documentation (#208)
  • Adding solvers in JuliaPro (#249)
  • Move JuliaPro and registry instructions to docs (#320)
  • Use POMDPs to solve a simple routine economics problem. (#326)
  • Link to Tiger example Down (#347)
  • Move to GitHub Actions CI (#348)
  • POMDPs.generate_s, POMDPs.generate_sr... and POMPDs.gen, PIMPDs.sampletype. (#349)
  • Leading the documentation with distributions (#355)
  • Package Maintenance Github Actions (#356)
  • Add DOCUMENTER_KEY ssh keys to all repos that use CompatHelper (#364)
  • LoadError: LoadError: UndefVarError: n_actions not defined (#365)
  • FAQ: What happens if I don't use the rng (#374)
  • return only one dimension in multidimensional observation? (#377)
  • Point to SymbolicMDPs.jl for PDDL support (#379)

Merged pull requests:

  • MassInstallAction: Install the TagBot workflow on this repository (#358) (@zsunberg)
  • removed travis (#359) (@zsunberg)
  • MassInstallAction: Install the CompatHelper workflow on this repository (#361) (@zsunberg)
  • MassInstallAction: Install the CompatHelper workflow on this repository (#362) (@zsunberg)
  • Updated build/coverage statuses in README (#371) (@mossr)
  • Update broken link (#372) (@potatoboiler)
  • Docs twostage (#375) (@zsunberg)
  • FAQ about what happens if you don't use the rng (#376) (@zsunberg)
  • added FAQ about the difference between gen and transition (#380) (@zsunberg)
  • CompatHelper: bump compat for NamedTupleTools to 0.14, (keep existing compat) (#381) (@github-actions[bot])

v0.9.3+docs3

2 years ago

Added documentation about transition, gen, and @gen

v0.9.3+docs

2 years ago

Updated documentation with a QuickPOMDPs- and distributions-centric guide.

v0.9.3

3 years ago

POMDPs v0.9.3

Diff since v0.9.2

Closed issues:

  • tag for distributions 0.24 compat (#327)
  • Github Discussions (#329)
  • stateindex for MDP without CartesianIndices (#332)
  • Solving "Battleship" with POMDPS.jl/BasicPOMCP (legal action spaces that change overtime) (#335)
  • ERROR: BoundsError: attempt to access 3-element Array{Float64,1} at index [4] (#340)

Merged pull requests:

  • CompatHelper: bump compat for "Distributions" to "0.25" (#346) (@github-actions[bot])

v0.9.2

3 years ago

POMDPs v0.9.2

Diff since v0.9.1

Closed issues:

  • JuliaPOMDP Registration Procedure (#271)
  • The main QuickPOMDPs interface works well (#299)
  • Adding solvers to Julia general registry (#303)
  • Solver package update checklist (#312)
  • Solver General Registration Checklist (#318)
  • Why is POMDPs v0.9 lower bounded to Julia v1.1? (#321)

Merged pull requests:

  • Add PointBasedValueIteration solver (#317) (@dominikstrb)
  • No registry instructions (#322) (@zsunberg)
  • CompatHelper: bump compat for "Distributions" to "0.24" (#323) (@github-actions[bot])

v0.9.1

3 years ago

POMDPs v0.9.1

Diff since v0.9.0

Closed issues:

  • The alternative to gen that keeps me up at night (#269)
  • How to fix the initial state? (#316)

Merged pull requests:

  • Code of Conduct (#314) (@zsunberg)
  • Contribution Instructions (#315) (@zsunberg)

v0.9.0

3 years ago

POMDPs v0.9.0

Diff since v0.8.4

Closed issues:

  • Focusing on POMDPs makes it harder to jump in and implement an MDP (#150)
  • updater=nothing means pass o or the initial belief/missing (#189)
  • Belief vector representation (#233)
  • Docs: only reward(m, s, a) shows up (#235)
  • Consider moving initialstate_distribution to the common section in the documentation (#237)
  • Remove dependency on LiBGit2? (#250)
  • DBN Construction Syntax (#261)
  • classic/traditional mdp and pomdp DBNs (#262)
  • The future of requirements and implemented (#263)
  • DDN Input nodes (#266)
  • Type stability issue in julia 1.0 (#279)
  • Update documentation for defining a solver (#280)
  • Transition function error (#285)
  • Fix DDN Structure mistake and move to SequentialDecisionBase.jl (#292)
  • Version 0.9 Meta Issue (#297)
  • 0.8 Lessons Learned (#298)
  • DiscreteExplicit(PO)MDP is not very useful (#300)
  • Syntactic sugar for gen(DDNOut(...), ...) (#302)
  • Remove customizable DDN in POMDPs.jl v0.9 (#306)
  • Consistent distribution semantics (change initialstate? change action?) (#308)
  • Role of gen and @gen (#309)
  • Issues with rolling out v0.9 (#310)
  • Some problem (#311)

Merged pull requests:

  • Version 0.9 (#307) (@zsunberg)