Evolutionary Algorithms: The Key to Solving Complex Java Puzzles

Passion for Technologyb.knopper@ncim.nl @BWKnopper
EVOLUTIONARY ALGORITHMS
The key to solving complex Java puzzles
JavaOne [BOF2913]
#javaone #EvolutionaryAlgorithms @BWKnopper

Let me introduce myself…
• Bas W. Knopper
• Dutch
• Java Developer
• NCIM (IT Services Company)
• 9 years Java exp
• AI enthousiast
• Soft spot for Evolutionary Algorithms
Slide 2 of 68

Questions
• A lot of ground to cover in 45 mins.
• Ask questions anytime!
• But don’t take it personal if I have to move on…
• We’ll discuss it over a <insert beverage of choice>!
• If time permits it
• Question round at the end as well
Slide 3 of 68

What I would like to accomplish…
• Interest
• Understanding
• How & when
• Add to toolbox
Slide 4 of 68

NASA
• Space Technology 5 mission
• launched March 22, 2006, and completed June 20, 2006
• Three full service 25-kilogram-class spacecraft
Slide 5 of 68

NASA Continued
• Needs even smaller antenna
• That still functions according to spec
• Need for solution that’s not easy to engineer
• So they used an EA that made these:
Slide 6 of 68

Slide 7 of 68

Recap
• Powerful example
• Evolving object
• First evolved object to fly in space
• How?
Slide 8 of 68

Evolution - “Survival of the fittest”
Finite
Resources
Lifeforms
with a basic
instinct
towards
Reproduction
Natural
Selection
Slide 9 of 68

Recombination
Mutation
Slide 10 of 68

Recombination
Mutation
Slide 11 of 68

From evolution to problem solvingEnvironment Problem
Individual Candidate Solution
Slide 12 of 68

Puzzle solving time!
• More down to earth example
• Travelling Salesman Problem
Slide 13 of 68

Travelling Salesman Problem
• Given n cities
• n = number of cities to visit
• Find (optimal) route for visiting all cities
• Visit every city only once
• Return to origin city
• Search space is huge
• For 30 cities there are 30! 10^32 possible routes
That’s 100.000.000.000.000.000.000.000.000.000.000 possible routes!
Brute force might not be the best solution…
Slide 14 of 68

Puzzle solving time!
• More down to earth example
• Travelling Salesman Problem
• Use case to show you
• Evolutionary Algorithm Design
• Plain Java Code
• Demo
Slide 15 of 68

Different EA’s
• Genetic Algorithms (GA’s)
• Evolution Strategies (ES’s)
• Evolutionary Programming (EP)
• Genetic Programming (GP)
• All have the same basis
Slide 16 of 68

EA Algorithm (Pseudocode)
INITIALISE population with random candidate solutions;
EVALUATE each candidate;
WHILE ( TERMINATION CONDITION is not satisfied ) {
1 SELECT parents;
2 RECOMBINE pairs of parents;
3 MUTATE the resulting offspring;
4 EVALUATE new candidates;
5 SELECT individuals for the next generation;
}
Slide 17 of 68

1 SELECT parents;
}
Slide 18 of 68

Candidate Solution - Representation
• n = 6
• And base city 1
• Label cities 1,2,3,4,5,6
• Candidate Solution
• Signifying the route
• for n = 10
1563421
110947356821
Slide 19 of 68

Recombination
Mutation
1653421 1653421
1653421
1653421
1653421
Slide 20 of 68

1 SELECT parents;
}
Slide 21 of 68

1 SELECT parents;
}
Slide 22 of 68

Evaluation Function (Fitness Function)
• Summed distance:
• d1,2 + d2,4 + … + d5,1
• Minimize!
1563421
Slide 23 of 68

/**
* Calculates the total distance of the whole route and stores it as this
* candidate solution's fitness
*/
private void calculateFitness() {
/* initialize total distance */
double totalDistance = 0;
/*
* For all Cities in the route (except the last one) get the distance between this
* City and the next and add it to the totalDistance
*/
for (int i = 0; i < route.size() - 1; i++) {
City city = route.get(i);
City nextCity = route.get(i + 1);
totalDistance += city.calculateDistance(nextCity);
}
/* store totalDistance as this candidate solution's fitness */
this.fitness = totalDistance;
}
Slide 24 of 68

/**
*/
/*
*/
}
}
Slide 25 of 68

/**
*/
/*
*/
}
}
Slide 26 of 68

/**
*/
/*
*/
}
}
Slide 27 of 68

/**
*/
/*
*/
}
}
Slide 28 of 68

/**
*/
/*
*/
}
}
Slide 29 of 68

/**
*/
/*
*/
}
}
Slide 30 of 68

1 SELECT parents;
}
Slide 31 of 68

Termination Condition
• EA’s are stochastic
• May never find optimum
• Combination
• Max number of runs (generations)
• Sure to terminate
• Either:
• Fitness threshold
• Fitness improvement remains under a threshold over runs
• Population diversity drops under a certain threshold
Slide 32 of 68

1 SELECT parents;
}
Slide 33 of 68

Parent Selection
• Where x is the number of parents:
• Pick x best
• Random x
• Best x out of random y
• In our example:
• Best x out of random y
Slide 34 of 68

private List<CandidateSolution> parentSelection() {
List<CandidateSolution> tempPopulation = new ArrayList<CandidateSolution>(population);
List<CandidateSolution> randomCandidates = new ArrayList<CandidateSolution>();
/* create parent pool */
for(int i = 0; i < parentPoolSize; i++)
{
/* select a random candidate solution from the temp population */
int randomlySelectedIndex = random.nextInt(tempPopulation.size());
CandidateSolution randomSelection = tempPopulation.get(randomlySelectedIndex);
randomCandidates.add(randomSelection);
/* delete the candidate from the temp population, so we can't pick it again */
tempPopulation.remove(randomlySelectedIndex);
}
/* Sort the population so that the best candidates are up front */
Collections.sort(randomCandidates);
/* return a list with size parentSelectionSize with the best CandidateSolutions */
return randomCandidates.subList(0, parentSelectionSize);
}
Slide 35 of 68

{
}
}
Slide 36 of 68

{
}
}
Slide 37 of 68

{
}
}
Slide 38 of 68

{
}
}
Slide 39 of 68

{
}
}
Slide 40 of 68

{
}
}
Slide 41 of 68

{
}
}
Slide 42 of 68

1 SELECT parents;
}
Slide 43 of 68

Recombination
• In our example:
• half-half does not work
• “Cut-and-crossfill”
1653421
1421
1245631
15421 135421 1635421 1425631
1245631 1245631 1245631
Slide 44 of 68

public List<CandidateSolution> recombine(CandidateSolution otherParent) {
/* get routes of both parents */
List<City> parentRoute1 = getRoute();
List<City> parentRoute2 = otherParent.getRoute();
/* initialize the routes for the children */
List<City> childRoute1 = new ArrayList<City>();
/* randomize cutIndex for "cross-and-fill point" */
int cutIndex = new Random().nextInt(parentRoute1.size());
/* copy the first part of the parents cut into the children */
childRoute1.addAll(parentRoute1.subList(0, cutIndex));
/* perform crossfill for both children */
crossFill(childRoute1, parentRoute2, cutIndex);
/* create new children using the new children routes */
CandidateSolution child1 = new CandidateSolution(childRoute1);
/* put the children in a list and return it (omitted for layout reasons) */
}
Slide 45 of 68

}
Slide 46 of 68

}
Slide 47 of 68

}
Slide 48 of 68

}
Slide 49 of 68

}
Slide 50 of 68

/**
* Check the rest of the route in the crossing parent and add the cities
* that are not yet in the child (in the order of the route of the crossing
* parent)
*/
private void crossFill(List<City> childRoute, List<City> parentRoute,
int cutIndex) {
/* traverse the parent route from the cut index on and add every city not yet in the child to the child */
for (int i = cutIndex; i < parentRoute.size(); i++) {
City nextCityOnRoute = parentRoute.get(i);
if (!childRoute.contains(nextCityOnRoute)) {
childRoute.add(nextCityOnRoute);
}
}
/* traverse the parent route from the start of the route and add every city not yet in the child to the child */
for (int i = 0; i < cutIndex; i++) {
City nextCityOnRoute = parentRoute.get(i);
if (!childRoute.contains(nextCityOnRoute)) {
childRoute.add(nextCityOnRoute);
}
}
}
1425631
1425631
Slide 51 of 68

1 SELECT parents;
}
Slide 52 of 68

1653421
Mutation
• Change of nr won’t work
• Infeasible candidate solution
• Swap to the rescue!
1653421 1253461
Slide 53 of 68

/**
* Mutates the current individual by swapping two random cities in its
* route.
*/
public void mutate() { 1653421
1653421
1253461
Random random = new Random();
/* randomly select two indices in the route */
int indexFirstCity = random.nextInt(route.size());
int indexSecondCity = random.nextInt(route.size());
/* Make sure they are different */
while (indexFirstCity == indexSecondCity) {
indexSecondCity = random.nextInt(route.size());
}
/* retrieve the Cities on the given indices */
City firstCity = route.get(indexFirstCity);
City secondCity = route.get(indexSecondCity);
/* Changer! */
route.set(indexFirstCity, secondCity);
route.set(indexSecondCity, firstCity);
}
Slide 54 of 68

1 SELECT parents;
}
Slide 55 of 68

1 SELECT parents;
}
Slide 56 of 68

Survivor Selection
• Replacement Strategy
• Replace Worst
Slide 57 of 68

/**
* Selects the survivors by removing the worst candidate
* solutions from the list, so we have the original
* population size again
*/
private void selectSurvivors() {
Collections.sort(population);
population = population.subList(0, populationSize);
}
Slide 58 of 68

Tuning…
• Mutation probability
• Population size
• Nr of offspring
• Termination condition (# runs or fitness)
• Parent selection
• Survival selection
• Initialisation
• Random
Slide 59 of 68

EA Behavior
Figures from “Introduction to Evolutionary Computing” by A.E. Eiben & J.E. Smith (Springer)
Slide 60 of 68

Demo!
• Used NASA World Wind to map route on 
• https://github.com/ncim-tsp/JavaOneTSP.git
Slide 61 of 68

Java Frameworks & API’s
http://jgap.sourceforge.net/http://watchmaker.uncommons.org
Slide 62 of 68

Java Frameworks & API’s
• ECJ
• http://cs.gmu.edu/~eclab/projects/ecj/
• MOEA Framework
• http://www.moeaframework.org
• JEAF
• https://github.com/GII/JEAF
• …
Slide 63 of 68

With great power comes great responsibility
• Can I find a solution using a brute-force approach?
• (within a reasonable amount of time)
• Am I facing an optimization or search problem?
• Can I encode a candidate solution to the problem?
• Representation possible?
• Can I determine the fitness of a candidate solution?
Slide 64 of 68

than a fool can learn from a wise answer”
– Bruce Lee
“A wise man can learn more from a foolish question
Slide 65 of 68

Additional questions?
• Contact me on @BWKnopper
• Google it!
• There’s lots to find…
• Papers
• Demo’s
• Aforementioned Frameworks/API’s
Slide 66 of 68

Have a great week @JavaOne
That’s all folks!
Thank you.
Slide 67 of 68

Evolutionary Algorithms: The Key to Solving Complex Java Puzzles

Recommended

More Related Content

Similar to Evolutionary Algorithms: The Key to Solving Complex Java Puzzles (20)

Recently uploaded (20)

Evolutionary Algorithms: The Key to Solving Complex Java Puzzles