Soft Computing Library

A Java library implementing Genetic Algorithms, Fuzzy Logic, and Neural Networks with practical case studies.

Overview

Modular implementations of three core soft computing techniques with real-world applications:

• 🧬 Genetic Algorithms: Complete framework with multiple selection, crossover, mutation strategies
• 🧠 Fuzzy Logic: Mamdani/Sugeno inference with customizable rules and membership functions
• 🤖 Neural Networks: Feedforward networks with backpropagation and multiple optimizers

Quick Start

# Compile
javac -d bin src/**/*.java
# Run
java -cp bin Main

Select from three demo applications:

1. Job Scheduling - GA optimization
2. Restaurant Tipping - Fuzzy logic system
3. Facility Classification - Neural network

Usage Examples

Genetic Algorithm

GeneticAlgorithm<Integer> ga = newGeneticAlgorithm<>();
ga.setPopulationSize(50);
ga.setGenerations(100);
ga.setCrossoverRate(0.7);
ga.setMutationRate(0.01);
ga.setChromosome(newIntChromosome(10, 0, 100));
ga.setFitnessFunction(chromosome -> /* fitness logic */);
ga.setSelectionStrategy(newTournamentSelection<>(3));
ga.setCrossoverStrategy(newUniformCrossover<>());
ga.setMutationStrategy(newSwapMutation<>());
ga.run();

Available Strategies:

• Selection: Tournament, Roulette Wheel, Rank
• Crossover: Single-point, N-point, Uniform
• Mutation: Swap, BitFlip, Scramble
• Replacement: Elitist, Generational, Steady-state

Fuzzy Logic

// Define variables and fuzzy setsLinguisticVariableservice = newLinguisticVariable("service", 0, 10);
service.addFuzzySet(newFuzzySet("poor", newTriangular(0, 0, 5)));
service.addFuzzySet(newFuzzySet("excellent", newTriangular(5, 10, 10)));
LinguisticVariabletip = newLinguisticVariable("tip", 0, 25);
tip.addFuzzySet(newFuzzySet("low", newTriangular(0, 0, 13)));
tip.addFuzzySet(newFuzzySet("high", newTriangular(13, 25, 25)));
// Create rulesRuleBaseruleBase = newRuleBase();
ruleBase.addRule(newFuzzyRule(
List.of(newFuzzyRule.Condition("service", "poor")), "tip", "low"));
// EvaluateFuzzyLogicSystemfls = newFuzzyLogicSystem(List.of(service), tip, ruleBase);
doubleresult = fls.evaluate(Map.of("service", 7.5));

Features: Triangular/Trapezoidal membership, Mamdani/Sugeno inference, Centroid defuzzification

Neural Network

Hyperparametersparams = newHyperparameters.Builder()
.layerSizes(newint[]{4, 10, 3})
.activationFunction(newReLU())
.lossFunction(newCrossEntropy())
.optimizer(newAdam(0.001))
.epochs(100)
.batchSize(32)
.build();
NeuralNetworknn = newNeuralNetwork(params);
nn.train(trainingData, labels);
double[] prediction = nn.predict(input);

Components: Sigmoid/ReLU/Tanh activation, MSE/Cross-entropy loss, SGD/Adam/RMSprop optimizers

Project Structure

src/
├── Main.java # Entry point
├── CaseStudies/
│ ├── JobScheduling/ # GA: Job scheduling
│ ├── RestaurantTipping/ # FL: Tipping system
│ └── FacilityClassification/ # NN: Classification
└── SoftComputingTechniques/
├── ga/ # Genetic algorithms
├── fl/ # Fuzzy logic
└── nn/ # Neural networks

Key Components

Module	Core Classes
GA	GeneticAlgorithm, Chromosome, FitnessFunction, Selection, CrossOver, Mutation
FL	FuzzyLogicSystem, LinguisticVariable, FuzzySet, RuleBase, InferenceEngine
NN	NeuralNetwork, Layer, Hyperparameters, ActivationFunction, Optimizer

Case Studies

Job Scheduling (GA) - Optimize job scheduling on machines with time constraints
Restaurant Tipping (FL) - Calculate tips based on service/food quality
Facility Classification (NN) - Classify facilities by accessibility features