Fast Potts Models in Python

PottsPlayground is a Python package for constructing and simulating combinatorial optimization problems represented as Potts Models. It’s purpose is twofold: providing an easily readable interface for constructing Potts models, and providing a fast C++/CUDA backend for annealing or sampling from models.

Several example combinatorial problems are built-in, ranging from simple graph coloring to logic element placement for Ice40 FPGAs. There are also methods for converting any Potts model into an Ising model, and for creating minor-embedded representations of Ising models.

• Getting Started
  • Installation
  • Using Built-in Tasks
  • Creating a New Task
• Potts Model Constructor
  • PottsModel
    • PottsModel.AddBias()
    • PottsModel.AddKernel()
    • PottsModel.AddSpins()
    • PottsModel.AddWeight()
    • PottsModel.Compile()
    • PottsModel.EnergyBands()
    • PottsModel.EvalCost()
    • PottsModel.EvalDDE()
    • PottsModel.EvalDE()
    • PottsModel.EvalPE()
    • PottsModel.GetSpinBias()
    • PottsModel.GetSpinFromState()
    • PottsModel.IndexOf()
    • PottsModel.InitKernelManager()
    • PottsModel.ListSpins()
    • PottsModel.PinSpin()
    • PottsModel.SetBias()
    • PottsModel.SetSpinInState()
    • PottsModel.SpinSize()
    • PottsModel.TotalWeight()
    • PottsModel.__init__()
• Built-in Tasks
  • Graph Coloring
    • GraphColoring
      • GraphColoring.DisplayState()
      • GraphColoring.__init__()
  • Traveling Salesman
    • TravelingSalesman
      • TravelingSalesman.DisplayState()
      • TravelingSalesman.__init__()
• Model Converters
  • Binarized
    • Binarized
      • Binarized.IsingToPottsSemantics()
      • Binarized.PottsToIsingSemantics()
      • Binarized.__init__()
  • MinorEmbedded
    • MinorEmbedded
      • MinorEmbedded.FuzzyMinorToState()
      • MinorEmbedded.StateToMinor()
      • MinorEmbedded.__init__()
• Running a Model
  • Annealing API
    • PottsPlayground.Anneal()
  • Models
    • model=”Potts”
    • model=”PottsPrecompute”
    • model=”Ising”
    • model=”IsingPrecompute”
    • model=”Swap”
    • model=”Swap+”
  • Parallelism
    • Replicates (nReplicates)
    • Multithreading (nWorkers)
    • Algorithmic
  • Algorithms
    • algo=”Simple” (CPU only)
    • algo=”ParallelTrials”
    • algo=”KMC” (CPU only)
  • Reports
• Annealing Schedules
  • LinearTemp()
  • SawtoothTempLog2Space()
  • constTemp()
  • constTempLog2Space()
• Weight Kernels
  • BinaryQuadratic()
  • Identity()
  • IsingSpin()