Skip to content

Latest commit

 

History

History
executable file
·
99 lines (76 loc) · 2.95 KB

OLD_README.md

File metadata and controls

executable file
·
99 lines (76 loc) · 2.95 KB
Raw

Learning simulations with ScaFi-Alchemist

  • ScaFi is an aggregate programming framework
  • Alchemist is a (meta-)simulator

Alchemist

Conceptual model

  • Molecule: name of a data item
  • Concentration: value associated to a particular molecule
  • Node: a container of molecules/reactions, living inside an environment
  • Environment: the Alchemist abstration for the space.
    • It is a container for nodes, and it is able to tell: a) Where the nodes are in the space - i.e. their position b) How distant are two nodes c) Optionally, it may provide support for moving nodes
  • Linking rule: a function of the current status of the environment that associates to each node a neighborhood
    • Neighborhood: an entity composed by a node (centre) + a set of nodes (neighbors)
  • Reaction: any event that can change (through an action) the state of the environment
    • Consists of 0+ conditions, 1+ actions, and a time distribution
    • Conditions, time distribution, static rate, and rate equation affect the frequency of the reaction
  • Alchemist implements an optimised version (NRM) of Gillespie's Stochastic Simulation Algorithm (SSA)

So

  • The system state depends on the configuration of molecules floating in it
  • The system evolution depends on the kinds of chemical reactions applicable over time

Another key concept is the dependency graph

  • Actions are outputs
  • Conditions are inputs

Running simulations

Use task configured in build.gradle.kts:

$ ./gradlew hello

Basically, all you need to do to launch a simulation (batch) is running it.unibo.alchemist.Alchemist with your simulation's descriptor. So, another approach is the following:

./gradlew fatJar # or shadowJar

java -Xmx2524m -cp "build/libs/<MYFATJAR>.jar" \
  it.unibo.alchemist.Alchemist \
  -b -var random \
  -y <path-to-yaml> -e data/20191004-test \
  -t 100 -p 1 -v &> exec.txt &

Scafi

The simulation descriptor must indicate incarnation: scafi and configure a RunScafiProgram action pointing to the class of some class implementing AggregateProgram.

incarnation: scafi

pools:
  - pool: &program
    - time-distribution:
        type: ExponentialTime
        parameters: [1]
      type: Event
      actions:
        - type: RunScafiProgram
          parameters: [it.unibo.casestudy.HelloWorld, 5.0] # second argument is retention time
    - program: send

An example ScaFi program is the following

package it.unibo.casestudy

import it.unibo.alchemist.model.implementations.molecules.SimpleMolecule
import it.unibo.alchemist.model.scafi.ScafiIncarnationForAlchemist._

class HelloWorld extends AggregateProgram with StandardSensors with Gradients {
  override def main(): Any = {
    val x = node.get[Int]("prova")
    node.put("prova2", x+1)
    classicGradient(mid==100)
  }
}

Plotting

$ python2.7 ./plotter.py plot.yml data/ .*-g.* myg
$ ./showImgs.sh data/imgs/ myg 2 tot