/ilc1.tsp

Primary LanguageJava

TP 4A 2020-21: The traveling salesman problem (TSP) with genetic algorithm

Questions

  1. Fork this github repository (2 points)
  2. Run the project
  3. Comment the 3 classes: City, Population & Tour (3 points)
  4. Commit your changes and check-it into github
  5. On github, make a pull request (1 point)
  6. Complete the next section: "Analyze IADT functions" into this file (2 points)
  7. Commit your changes (commit README.md)
  8. On github, update your pull request (1 point)
  9. Make tests defined into section: "Functions to test now" (call your test functions using the rule: GIVEN_WHEN_THEN) (10 points)
  10. Check the test coverage with the Jacoco coverage report provided into: /target/site/jacoco/index.html
  11. Commit your changes
  12. On github, update your pull request (1 point)
  13. that ends

Analyze IADT functions

Hereafter, for each classes give function to be tested. Remember that all functions does not have to be tested, some code can be checked by a simple inspection or an analyse, another code can be validated by definition (rules, etc) the others function can be tecsted by unit tests, integration tests or validation tests (IADT: Inspect, Analyse, Define, Test).

Validation Mode Why ? When ?
I: Inspect No test needed, just a simple static test (visual) on a subset of classes (commentaries, indentation, etc.)
A: Analyze Not easy to test but the algorithm is well know and can be check by a static analyze
D: Def/Dem You can Demonstrate the function or by Definition the function needs not test (generate code, call functions, etc.)
T: Test the function can be easily test

Class City

Function I A D T Comment
distanceToCity
getX
getY
toString

Class Tour

Function I A D T Comment
containsCity
getCity
getDistance
getFitness
setCity
tourSize
toString

Class Population

Function I A D T Comment
getFittest
getTour
populationSize
saveTour

Functions to test now

Class City

public double distanceTo(City city)

Class Tour

public int getDistance()

Class GA

public static Population evolvePopulation(Population pop)

Tricks & tips

  • Start with simple classes (models, libraries, etc) at the end make test into complex classes (classes with multiple relations)
  • Never test auto generate code
  • Never test creator (its a design problem)
  • Use @Test before each testFunction !
  • You can test a function than throws an exception with the argument: expected
@Test(expected=Exception.class)

Documentation

http://www.theprojectspot.com/tutorial-post/applying-a-genetic-algorithm-to-the-travelling-salesman-problem/5