Failover error?

I am struggling to implement Resilient Propagation. I have already implemented the backpropagation algorithm for training the neural network and it works as expected for the XOR-Net i.e. It takes about 600 Epochs to reset the error below 1%. Now I have tried implementing Resilient Propagation ( http://en.wikipedia.org/wiki/Rprop ) for the same problem and for the first few Epochs the Error drops quickly to 23% but then goes up to 50% and stays there.
I implemented it exactly as described at http://www.heatonresearch.com/book/introduction-neural-network-math.htmlbut this is a perplexing description: it differs from the wikipedia Rprop-Page AND from the implementation, as far as I know it was written by the same author as the book. I've also tried different implementations from different sources already, but nothing worked. Some differences between different sources:

• using signum (currentPartialDerivative) instead of signum (currentPartialDerivative * previousPartialDerivative) to calculate weight changes
• using the last weight change instead of the new update value for the new weight change
• Calculating weight change first and new update values ​​second

What's the correct way to implement this? Subsequently Implementation according to the book:

public ResilientPropagation() {
    initialUpdateValue = 0.01;
    deltaMaximum = 50;
    deltaMinimum = 0.000001;
    negativeEta = 0.5;
    positiveEta = 1.2;
    double zeroTolerance = 0.0000000000000001;
    signum = new Signum(zeroTolerance);

    init();
}

@Override
public double calculateWeightChange(Synapse synapse, double partialDerivative) {
    if (!synapseValues.containsKey(synapse)){
        double initialPartialDerivative = 0;
        synapseValues.put(synapse, new SynapseValues(initialUpdateValue, initialPartialDerivative));
    }

    SynapseValues values = synapseValues.get(synapse);
    double signChange = signum.value(values.lastPartialDerivative * partialDerivative);
    values.lastPartialDerivative = partialDerivative;
    double weightChange = 0;
    if (signChange > 0){
        newUpdateValue = Math.min(positiveEta * values.updateValue, deltaMaximum);
        weightChange = -1*newUpdateValue;
    } else if (signChange < 0){
        newUpdateValue = Math.max(negativeEta * values.updateValue, deltaMinimum);
        weightChange = newUpdateValue;
    } else {
        newUpdateValue = values.updateValue;
        double weightChange = 0;
    }
    values.updateValue = newUpdateValue;
    return weightChange;
}

      

        
        
        
      

    

The same method for normal Backpropagation works great:

@Override
public double calculateWeightChange(Synapse synapse, double partialDerivative) {
    double previousChange = previousWeightChange.get(synapse) != null ? previousWeightChange.get(synapse) : 0;
    double weightChange = learningRate * partialDerivative + momentum * previousChange;
    previousWeightChange.put(synapse, weightChange);
    return weightChange;
}