@@ -32,3 +32,5 @@ All the accepted submissions will be part of the market place and will be compil
-[SimPopLocal model](simpoplocal): a geographical model calibrated using genetic algorithms.
- [Sensitivity-Screening analysis](sensitivity/morris): a method to quickly analyze which inputs are influential on large spaces of parameters.
- [Global Sensitivity Analysis](sensitivity/saltelli): a variance based sensitivity analysis of model output.
-[Test Functions for NSGA2](nsga2-test-functions): reference functions to double check the correctedness of the NSGA2 algorithm, and also view examples of usage of NSGA2
Metaheuristics for optimization such as genetic algorithms leave a huge freedom to developers for their implementation.
Genetic Algorithms such as NSGA2 are even less trivial.
Their implementation in the [mgo](https://github.com/openmole/mgo) used in [OpenMole](https://openmole.org/) was adapted so they are efficient for computation on clusters or grids.
This leaves a lot of uncertainty to the user on what really happens.
Why should you, as a user, trust this implementation?
The answer is simple: you should not.
Any implementation in scientific computing should be verified.
## Test Functions
Many functions named "Test Functions" were proposed over time to test and compare multi-objective optimization algorithmls.
The [wikipedia page](https://en.wikipedia.org/wiki/Test_functions_for_optimization) lists several of them.
A test function is an optimization function which you can propose as the problem to optimize to your implementation.
Functions test:
- problems with multiple parameters and multiple objectives
- problems with constraints: a part of the space of parameter leads to computation errors as a result, so the algorithm should achieve to deal with it
- problems with parts of the space of solution that are tricky to detect, either because they are statistically unlikely to find, or because they are is a part of the space of solutions which is heavily constrained, etc.
Test tunctions enable the testing of several aspects including:
- coverage of the Pareto front,
- speed of convergence
Once a test function was explored by an implementation of an algorithm, you should compare the Pareto front obtained withou your implementation with the one expected in literature.
## Test and Learn
Test functions have another role: they constitute a simple example of optimization problem.
As a user, you can also learn to tune the parameters of the genetic algorithm as check when convergence
occurs.
## Test Workflows for OpenMole
We provide here a few examples of test functions which you can open with OpenMole.
For each workflow:
- Choose a test function.
- Run the workflow.
- Open the files with the solutions, check you understand them.
- Graph them and compare them with the literature.