Model Evaluation with OpenMOLE

Purpose

• Provide state-of-the-art exploration methods
• Tackle scalability using distributed computing
• Is not intrusive in the model logic
• Interoperate smoothly with your Models

1 - Model?

Your model here, with the plugins and data.

NetLogo on Fire

val model =
  NetLogo6Task(
    workDirectory / "Fire.nlogo",
    List("setup", "while [any? turtles] [go]")) set (
    inputs += seed,
    outputs += (seed, density),
    netLogoInputs += (density, "density"),
    netLogoOutputs += ("burned-trees", burned)
  )

Rrrrrrrrrr

val i = Val[Int]
  
val rTask = 
  RTask("""
    source("function.R")
    function(i)""") set (
    resources += workDirectory / "function.R",
    rInputs += i
  )

Feeling snaky? Python

val pyTask = 
  CARETask(
    workDirectory / "../python.bin",
    "python matrix.py data.csv 2 out.csv"
  ) 

2 - Method?

• Data reconstruction
• Parameter estimation
• Sensitivity analysis
• Robustness assessment
• Optimisation
• Diversity research
• Hybrid, ex: optimisation + diversity

Genetic Algorithm

Master/slave

Diversity research

3 - Execution environment?

Zero deployment approach

• User code is automatically deployed at runtime
• No prior knowledge of remote environment needed
• No installation required on any machine

Transparent access

• Access as the user would do it
• Use user credential
• No preliminary step

Automatic data transfers
(+ Replica management)

Today

Grid Computing

4000 cores (after a little administrative process... we can help you)

And now: hands on!

Useful Links

Thanks!

	romain.reuillon@iscpif.fr
	mathieu.leclaire@iscpif.fr
	j.passerat-palmbach@imperial.ac.uk