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).

Class City

Function	I A D T	Comment
distanceToCity	T	Easy to test
getX	D	Generated code
getY	D	Generated code
toString	I	No test needed, simple function

Class Tour

Function	I A D T	Comment
containsCity	D	Only one function called
getCity	I	No test needed, simple function
getDistance	T	can be tested and interesting to test
getFitness	I	No test needed because the function is a simple division
setCity	D	One function call and two simple instructions
tourSize	D	just call of one basic function
toString	I	No test needed, simple function

Class Population

Function	I A D T	Comment
getFittest	T	can be tested and interesting to test abd it's fundamental for the algorithm
getTour	D	just call of one basic function
populationSize	D	just call of one basic function
saveTour	D	just call of one basic function

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