Simulation of Artificial Neural Networks Gaming Project in C++

Projects Inventory

3 months ago

#include <iostream>
#include <vector>
#include <cmath>
#include <cstdlib>
#include <ctime>

using namespace std;

// Sigmoid activation function
double sigmoid(double x) {
    return 1.0 / (1.0 + exp(-x));
}

// Derivative of the sigmoid function
double sigmoidDerivative(double x) {
    return x * (1.0 - x);
}

// Neural network class
class NeuralNetwork {
private:
    vector<vector<double>> weights;
    vector<double> biases;
    vector<double> outputs;

public:
    NeuralNetwork(int inputSize, int outputSize) {
        // Initialize weights and biases
        weights.resize(outputSize, vector<double>(inputSize));
        biases.resize(outputSize);
        outputs.resize(outputSize);

        srand(static_cast<unsigned>(time(0))); // Seed for random number generation

        for (int i = 0; i < outputSize; ++i) {
            biases[i] = static_cast<double>(rand()) / RAND_MAX; // Random bias
            for (int j = 0; j < inputSize; ++j) {
                weights[i][j] = static_cast<double>(rand()) / RAND_MAX; // Random weight
            }
        }
    }

    // Forward propagation
    void forward(const vector<double>& inputs) {
        for (size_t i = 0; i < weights.size(); ++i) {
            double activation = biases[i];
            for (size_t j = 0; j < weights[i].size(); ++j) {
                activation += weights[i][j] * inputs[j];
            }
            outputs[i] = sigmoid(activation);
        }
    }

    // Print the network outputs
    void printOutputs() const {
        cout << "Outputs: ";
        for (const double& output : outputs) {
            cout << output << " ";
        }
        cout << endl;
    }
};

int main() {
    int inputSize = 3;
    int outputSize = 1;
    NeuralNetwork nn(inputSize, outputSize);

    vector<double> inputs = {0.5, 0.3, 0.2}; // Example inputs

    cout << "Running neural network simulation..." << endl;

    nn.forward(inputs);
    nn.printOutputs();

    return 0;
}

#include <iostream>

#include <vector>

#include <cmath>

#include <cstdlib>

#include <ctime>

using namespace std;

// Sigmoid activation function

double sigmoid(double x) {

return 1.0 / (1.0 + exp(-x));

}

// Derivative of the sigmoid function

double sigmoidDerivative(double x) {

return x * (1.0 - x);

}

// Neural network class

class NeuralNetwork {

private:

vector<vector<double>> weights;

vector<double> biases;

vector<double> outputs;

public:

NeuralNetwork(int inputSize, int outputSize) {

// Initialize weights and biases

weights.resize(outputSize, vector<double>(inputSize));

biases.resize(outputSize);

outputs.resize(outputSize);

srand(static_cast<unsigned>(time(0))); // Seed for random number generation

for (int i = 0; i < outputSize; ++i) {

biases[i] = static_cast<double>(rand()) / RAND_MAX; // Random bias

for (int j = 0; j < inputSize; ++j) {

weights[i][j] = static_cast<double>(rand()) / RAND_MAX; // Random weight

}

// Forward propagation

void forward(const vector<double>& inputs) {

for (size_t i = 0; i < weights.size(); ++i) {

double activation = biases[i];

for (size_t j = 0; j < weights[i].size(); ++j) {

activation += weights[i][j] * inputs[j];

}

outputs[i] = sigmoid(activation);

}

// Print the network outputs

void printOutputs() const {

cout << "Outputs: ";

for (const double& output : outputs) {

cout << output << " ";

}

cout << endl;

}

};

int main() {

int inputSize = 3;

int outputSize = 1;

NeuralNetwork nn(inputSize, outputSize);

vector<double> inputs = {0.5, 0.3, 0.2}; // Example inputs

cout << "Running neural network simulation..." << endl;

nn.forward(inputs);

nn.printOutputs();

return 0;

}

Explanation

Headers:
- <iostream>: For input and output operations.
- <vector>: For using the std::vector container.
- <cmath>: For mathematical functions like exp()
  
  Advertisement
  
  .
- <cstdlib>: For rand() and srand() functions.
- <ctime>: For time() function to seed the random number generator.
Activation Function:
- double sigmoid(double x): Computes the sigmoid activation function.
- double sigmoidDerivative(double x): Computes the derivative of the sigmoid function (not used in this simple example but essential for learning).
NeuralNetwork Class:
- Attributes:
  - weights: 2D vector holding the weights of the neurons.
  - biases: Vector holding the biases for each neuron.
  - outputs: Vector to store the output of each neuron.
- Constructor:
  - Initializes weights and biases with random values.
- Methods:
  - void forward(const vector<double>& inputs): Performs forward propagation by computing the weighted sum of inputs, adding the bias, and applying the sigmoid activation function.
  - void printOutputs() const: Prints the output of the network.
Main Function:
- Initialization:
  - Creates a NeuralNetwork object with a specified input size and output size.
  - Defines a vector of inputs.
- Simulation:
  - Runs forward propagation with the example inputs.
  - Prints the network outputs.

Notes:

Simple ANN: This example uses a single-layer neural network with randomly initialized weights and biases. It demonstrates basic forward propagation.

No Training: This program does not include training or backpropagation; it only shows how to perform forward propagation with a neural network.
Activation Function: The sigmoid function is used for activation, which is common in binary classification tasks.

Explanation

Related Posts: