polynomialNetworkTest.cpp

Go to the documentation of this file.

#include <gtest/gtest.h>
#include <mdp/learning/reinforcementLearning/actionValuesFunction/polynomialNetwork.h>

TEST(PolynomialNetworkTest, TestGetOutput)
{
    size_t I = 2;
    size_t O = 1;
    size_t H = 3;
    Mdp::PolynomialNetwork pol(nullptr, I, O, H);
    auto  inputWeights = std::vector<std::vector<double>>(I+1, std::vector<double>(H));
    auto outputWeights = std::vector<std::vector<double>>(H, std::vector<double>(O));
    inputWeights[0][0] = 1.0;
    inputWeights[0][1] = 1.0;
    inputWeights[0][2] = 1.0;
    inputWeights[1][0] = 1.0;
    inputWeights[1][1] = 1.0;
    inputWeights[1][2] = 1.0;
    inputWeights[2][0] = 1.0;
    inputWeights[2][1] = 1.0;
    inputWeights[2][2] = 1.0;

    outputWeights[0][0] = 1.0;
    outputWeights[1][0] = 1.0;
    outputWeights[2][0] = 1.0;

    pol.initializeWeights(inputWeights, outputWeights);

    std::vector<double> input{4.0, 5.0};

    std::vector<double> out = pol.getOutput(input);
    double actualOutput = out[0];
    std::vector<double> hidden = std::vector<double>(H, 0.0);
    
    double expectedOutput = 1.0 + 10.0 + 100.0;
    EXPECT_DOUBLE_EQ(expectedOutput, actualOutput);
}