/// <summary>
/// Creates a new solver.
/// </summary>
/// <param name="settings"></param>
/// <param name="selector"></param>
/// <param name="combined_operation"></param>
/// <param name="problem"></param>
/// <param name="fitness_calculator"></param>
public Solver(
ProblemType problem,
SolverSettings settings,
ISelector <GenomeType, ProblemType, WeightType> selector,
IOperation <GenomeType, ProblemType, WeightType> combined_operation,
IFitnessCalculator <GenomeType, ProblemType, WeightType> fitness_calculator)
{
_problem = problem;
_settings = settings;
_selector = selector;
_mutation_operation = combined_operation;
_crossOverOperation = combined_operation;
_generation_operation = combined_operation;
_fitness_calculator = fitness_calculator;
}
/// <summary>
/// Creates a new solver.
/// </summary>
/// <param name="problem"></param>
/// <param name="settings"></param>
/// <param name="selector"></param>
/// <param name="mutation"></param>
/// <param name="cross_over"></param>
/// <param name="generation"></param>
/// <param name="fitness_calculator"></param>
/// <param name="accept_only_better_cross_over"></param>
/// <param name="accept_only_better_mutation"></param>
public Solver(
ProblemType problem,
SolverSettings settings,
ISelector <GenomeType, ProblemType, WeightType> selector,
IMutationOperation <GenomeType, ProblemType, WeightType> mutation,
ICrossOverOperation <GenomeType, ProblemType, WeightType> cross_over,
IGenerationOperation <GenomeType, ProblemType, WeightType> generation,
IFitnessCalculator <GenomeType, ProblemType, WeightType> fitness_calculator,
bool accept_only_better_cross_over, bool accept_only_better_mutation)
{
_problem = problem;
_settings = settings;
_selector = selector;
_generation_operation = generation;
_mutation_operation = mutation;
_crossOverOperation = cross_over;
_fitness_calculator = fitness_calculator;
_accept_only_better_on_cross_over = accept_only_better_cross_over;
_accept_only_better_on_mutation = accept_only_better_mutation;
}
/// <summary>
/// Returns a solution found using best-placement.
/// </summary>
/// <returns></returns>
protected override IRoute DoSolve(OsmSharp.Math.TSP.Problems.IProblem problem)
{
// create the settings.
SolverSettings settings = new SolverSettings(
-1,
-1,
1000000000,
-1,
-1,
-1);
Solver<List<int>, GeneticProblem, Fitness> solver =
new Solver<List<int>, GeneticProblem, Fitness>(
new GeneticProblem(problem),
settings,
null,
null,
null,
_generation_operation,
new FitnessCalculator(),
true, false);
Population<List<int>, GeneticProblem, Fitness> population =
new Population<List<int>, GeneticProblem, Fitness>(true);
while (population.Count < _population_size)
{
// generate new.
Individual<List<int>, GeneticProblem, Fitness> new_individual =
_generation_operation.Generate(solver);
// add to population.
population.Add(new_individual);
}
// select each individual once.
Population<List<int>, GeneticProblem, Fitness> new_population =
new Population<List<int>, GeneticProblem, Fitness>(true);
Individual<List<int>, GeneticProblem, Fitness> best = null;
int stagnation = 0;
while (stagnation < _stagnation)
{
while (new_population.Count < _population_size)
{
// select an individual and the next one.
int idx = OsmSharp.Math.Random.StaticRandomGenerator.Get().Generate(population.Count);
Individual<List<int>, GeneticProblem, Fitness> individual1 = population[idx];
Individual<List<int>, GeneticProblem, Fitness> individual2 = null;
if (idx == population.Count - 1)
{
individual2 = population[0];
}
else
{
individual2 = population[idx + 1];
}
population.RemoveAt(idx);
Individual<List<int>, GeneticProblem, Fitness> new_individual = _cross_over_operation.CrossOver(solver,
individual1, individual2);
new_individual.CalculateFitness(solver.Problem, solver.FitnessCalculator);
if (new_individual.Fitness.CompareTo(
individual1.Fitness) < 0)
{
new_population.Add(new_individual);
}
else
{
new_population.Add(individual1);
}
}
population = new_population;
population.Sort(solver, solver.FitnessCalculator);
new_population = new Population<List<int>, GeneticProblem, Fitness>(true);
if (best == null ||
best.Fitness.CompareTo(population[0].Fitness) > 0)
{
stagnation = 0;
best = population[0];
}
else
{
stagnation++;
}
//// report progress.
//OsmSharp.IO.Output.OutputStreamHost.ReportProgress(stagnation,_stagnation,
// "OsmSharp.Math.TSP.EdgeAssemblyGenetic.EdgeAssemblyCrossOverSolver",
// "Solving using EAX...");
}
List<int> result = new List<int>(best.Genomes);
result.Insert(0, 0);
//return new SimpleAsymmetricRoute(result, true);
return DynamicAsymmetricRoute.CreateFrom(result);
}
/// <summary>
/// Solves the problem using a GA.
/// </summary>
/// <param name="problem"></param>
/// <returns></returns>
protected override IRoute DoSolve(OsmSharp.Math.TSP.Problems.IProblem problem)
{
//int population_size = 10;
//if (problem.Size < 100)
//{
// population_size = System.Math.Max(problem.Size * 3, 10);
// if (problem.Size < 10)
// {
// population_size = 1;
// }
//}
//if (problem.Size < 1000)
//{
// population_size = problem.Size / 4;
//}
// create the settings.
SolverSettings settings = new SolverSettings(
_stagnation_count,
_population,
1000000000,
_eltism,
_cross,
_mutation);
//List<IMutationOperation<List<int>, GeneticProblem, Fitness>> mutators = new List<IMutationOperation<int,GeneticProblem,Fitness>>();
////mutators.Add(new DefaultMutationOperation());
////mutators.Add(new BestPlacementMutationOperation());
//mutators.Add(new BestDetailedPlacementMutationOperation());
//List<double> probabilities = new List<double>();
//probabilities.Add(1);
////probabilities.Add(0.5);
////probabilities.Add(0.3);
//CombinedMutation<List<int>, GeneticProblem, Fitness> mutation = new CombinedMutation<int,GeneticProblem,Fitness>(
// StaticRandomGenerator.Get(),
// mutators,
// probabilities);
////SequentialContructiveCrossoverOperator cross_over = new SequentialContructiveCrossoverOperator();
//BestDetailedPlacementCrossOverOperation cross_over = new BestDetailedPlacementCrossOverOperation();
////BestPlacementCrossOverOperation cross_over = new BestPlacementCrossOverOperation();
////EdgeRecombinationCrossOverOperation cross_over = new EdgeRecombinationCrossOverOperation();
//BestPlacementGenerationOperation generation = new BestPlacementGenerationOperation();
////RandomGenerationOperation generation = new RandomGenerationOperation();
ISelector<List<int>, GeneticProblem, Fitness> selector = new RandomSelector<List<int>, GeneticProblem, Fitness>();
//ISelector<List<int>, GeneticProblem, Fitness> selector = new TournamentBasedSelector<List<int>, GeneticProblem, Fitness>(75, 0.01);
solver =
new Solver<List<int>, GeneticProblem, Fitness>(
new GeneticProblem(problem),
settings,
selector,
_mutation_operation,
_cross_over_operation,
_generation_operation,
new FitnessCalculator(),
true, false);
solver.NewFittest += new Solver<List<int>, GeneticProblem, Fitness>.NewFittestDelegate(solver_NewFittest);
solver.NewGeneration += new Solver<List<int>, GeneticProblem, Fitness>.NewGenerationDelegate(solver_NewGeneration);
List<int> result = new List<int>(solver.Start(null).Genomes);
result.Insert(0, 0);
return new SimpleAsymmetricRoute(result, true);
}
/// <summary>
/// Executes a solver procedure.
/// </summary>
/// <param name="problem"></param>
/// <returns></returns>
internal override MaxTimeSolution Solve(MaxTimeProblem problem)
{
// _customers = problem.Customers;
//float[] solutions = OsmSharp.Math.VRP.Core.BestPlacement.CheapestInsertionHelper.CalculateBestValues(
// problem, _customers);
generations = 0;
_max_generations = 10000000;
// calculate one tsp solution.
//Tools.Math.TSP.ISolver tsp_solver = new OsmSharp.Math.TSP.EdgeAssemblyGenetic.EdgeAssemblyCrossOverSolver(_population, _stagnation,
// new OsmSharp.Math.TSP.Genetic.Solver.Operations.Generation._3OptGenerationOperation(),
// new OsmSharp.Math.TSP.Genetic.Solver.Operations.CrossOver.EdgeAssemblyCrossover(30,
// OsmSharp.Math.TSP.Genetic.Solver.Operations.CrossOver.EdgeAssemblyCrossover.EdgeAssemblyCrossoverSelectionStrategyEnum.SingleRandom,
// true));
//IRoute tsp_solution = tsp_solver.Solve(new OsmSharp.Routing.VRP.NoDepot.MaxTime.TSPPlacement.TSPProblem(
// problem));
// initialize the generation.
IGenerationOperation<MaxTimeSolution, MaxTimeProblem, Fitness> generation =
//new SolverGenerationOperation(new TSPPlacement.TSPPlacementSolver<ResolvedType>(
// this.Router, this.Max, this.DeliveryTime, tsp_solution));
new OsmSharp.Routing.VRP.NoDepot.MaxTime.Genetic.Generation.RandomBestPlacement();
//new SolverGenerationOperation(new CheapestInsertionSolverWithImprovements<ResolvedType>(
// this.Router, this.Max, this.DeliveryTime, 5, 0.1f, true, 0.1f, false, 1f, null, null));
//new SolverGenerationOperation(new CheapestInsertionSolverWithImprovements<ResolvedType>(
// this.Router, this.Max, this.DeliveryTime, 5, 0.1f, true, 0.1f, true, 1f));
// initialize the crossover.
ICrossOverOperation<MaxTimeSolution, MaxTimeProblem, Fitness> cross_over =
new OsmSharp.Routing.VRP.NoDepot.MaxTime.Genetic.CrossOver.RouteExchangeOperation();
// initialize the mutation.
//IMutationOperation<MaxTimeSolution, MaxTimeProblem, Fitness> mutation =
// new VehicleMutation();
List<IMutationOperation<MaxTimeSolution, MaxTimeProblem, Fitness>> mutators =
new List<IMutationOperation<MaxTimeSolution, MaxTimeProblem, Fitness>>();
mutators.Add(new VehicleMutation());
//mutators.Add(new ThreeOptMutation());
//mutators.Add(new RedivideRouteMutation());
mutators.Add(new RoutePartExchangeMutation());
if (_probabilities == null)
{
_probabilities = new List<double>();
_probabilities.Add(0.2);
_probabilities.Add(0.6);
_probabilities.Add(0.2);
}
CombinedMutation<MaxTimeSolution, MaxTimeProblem, Fitness> mutation = new CombinedMutation<MaxTimeSolution, MaxTimeProblem, Fitness>(
StaticRandomGenerator.Get(),
mutators,
_probabilities);
SolverSettings settings = new SolverSettings(_stagnation, _population, _max_generations,
_elitism_percentage, _cross_percentage, _mutation_percentage);
MaxTimeProblem genetic_problem = problem;// new MaxTimeProblem(max, problem, solutions);
Solver<MaxTimeSolution, MaxTimeProblem, Fitness> solver =
new Solver<MaxTimeSolution, MaxTimeProblem, Fitness>(genetic_problem, settings,
new TournamentBasedSelector<MaxTimeSolution, MaxTimeProblem, Fitness>(5, 0.5),
mutation, //new ThreeOptMutation(),
cross_over, // new RouteExchangeOperation(), //new RouteExchangeOperation(), //new RouteExchangeAndVehicleOperation(), // Order1CrossoverOperation()
generation, //new RandomBestPlacement(),//new RandomGeneration(), //new RandomBestPlacement(),
new FitnessCalculator());
solver.NewFittest += new Solver<MaxTimeSolution, MaxTimeProblem, Fitness>.NewFittestDelegate(solver_NewFittest);
//solver.NewGeneration += new Solver<MaxTimeSolution, Problem, Fitness>.NewGenerationDelegate(solver_NewGeneration);
Individual<MaxTimeSolution, MaxTimeProblem, Fitness> solution = solver.Start(null);
//this.solver_NewFittest(solution);
MaxTimeSolution routes = solution.Genomes;
long ticks_after = DateTime.Now.Ticks;
StringBuilder sizes = new StringBuilder();
foreach (int size in routes.Sizes)
{
sizes.Append(size);
sizes.Append(" ");
}
StringBuilder weights = new StringBuilder();
foreach (double weight in solution.Fitness.Weights)
{
weights.Append(weight.ToString(CultureInfo.InvariantCulture));
weights.Append(" ");
}
return routes;
}
