Unity 2018 Artificial Intelligence Cookbook - Second Edition

How to do it...

We will implement the following steps in the same Bandit class we created before.

1. Define the required member variables:

float initialRegret = 10f;float[] regret;float[] chance;RPSAction lastOpponentAction;RPSAction[] lastActionRM;

2. Define the member function for initialization:

public void InitRegretMatching(){  if (init)    return;  // next steps}

3. Declare the local variables and initialize them:

numActions = System.Enum.GetNames(typeof(RPSAction)).Length;regret = new float[numActions];chance = new float[numActions];int i;for (i = 0; i < numActions; i++){  regret[i] = initialRegret;  chance[i] = 0f;}init = true;

4. Define the member function for computing the next action to be taken:

public RPSAction GetNextActionRM(){ // next ...