Substochastic Monte Carlo

Simulates a population of random walkers with birth-death resampling. qiotoolkit's implementation allows both the stepping probability alpha and the resampling factor beta to be specified as a schedule (over simulation steps).

Additionaly, you may specify the target_population and the number of steps_per_walker as input parametes.

Example

{
  "target": "substochastic-monte-carlo.qiotoolkit",
  "version": "1.0",
  "input_params": {
    "seed": 42,
    "alpha": {
      "initial_value": 0.8,
      "final_value": 0.2,
      "count": 1e3
    }
    "beta": {
      "initial_value": 0.2,
      "final_value": 0.8,
      "count": 1e3
    }
    "target_population": 800
  },
  "model": {...}
}

This simulates a population of 800 random walkers over the course of 200 steps. It starts out with an emphasis on stepping (alpha=0.8) and gradually changes to a resampling regime.

Note

Substochastic Monte-Carlo does currently NOT adjust the energy scale of your model. Therefore the magnitude of beta may need to be adjusted for your needs.

param	type	required/default	description
`seed`	integer	default: time-based
`target_population`	integer	required	The desired population size (throughout the simulation).
`alpha`	Schedule	required	How the stepping probability `alpha` should be changed over time.
`beta`	Schedule	required	How the resampling factor `beta` should be changed over time.
`steps_per_walker`	float >0		Number of steps to attempt for each walker.

{
  "$id": "solver/substochasticmontecarlo.qiotoolkit.schema",
  "$schema": "http://json-schema.org/draft-07/schema#",
  "title": "qiotoolkit Substochastic Monte Carlo",
  "type": "object",
  "required": ["target", "input_params"],
  "properties": {
    "target": {
      "type": "string",
      "const": "substochasticmontecarlo.qiotoolkit"
    },
    "input_params": {
      "type": "object",
      "required": ["target_population", "alpha", "beta"],
      "properties": {
        "target_population": {
          "type": "number",
          "minimum": 0,
          "multipleOf": 1
          "description": "The desired population size (throughout the simulation)."
        },
        "alpha": {
          "$ref": "../schedule/schedule.schema",
          "description": "How the stepping probability `alpha` should be changed over time."
        },
        "beta": {
          "$ref": "../schedule/schedule.schema",
          "description": "How the resampling factor `beta` should be changed over time."
        },
        "steps_per_walker" {
          "type": "integer",
          "description": "Number of steps to attempt for each walker.",
          "minimum": 1
        }
      }
    }
  }
}