public static void Run()
{
var world = new GridWorld();
var randomPolicy = new UniformRandomGridWorldPolicy();
var rewarder = new NegativeAtNonTerminalStatesGridWorldRewarder();
var values = new GridWorldValueTable(world);
// manually iterate a couple of times - optimal policy is greedy wrt
// initial random policy values
values.Evaluate(randomPolicy, rewarder);
values.Print();
var greedyPolicy = GreedyGridWorldPolicy.Create(world, values, rewarder);
values.Evaluate(greedyPolicy, rewarder);
values.Print();
greedyPolicy = GreedyGridWorldPolicy.Create(world, values, rewarder);
values.Evaluate(greedyPolicy, rewarder);
values.Print();
greedyPolicy.Print();
}
public static void Run()
{
var world = new GridWorld();
var policy = new UniformRandomGridWorldPolicy();
var rewarder = new NegativeAtNonTerminalStatesGridWorldRewarder();
var gridValues = new GridWorldValueTable(world);
gridValues.Evaluate(policy, rewarder);
gridValues.Print();
}
public static void Evaluates_to_same_values_as_gridworld_value_table()
{
var gridWorld = new GridWorld();
var rewarder = new NegativeAtNonTerminalStatesGridWorldRewarder();
var policy = new UniformRandomGridWorldPolicy();
var gridValues = new GridWorldValueTable(gridWorld);
var genericValues = new ValueTable <GridWorldState, GridWorldAction>(gridWorld);
gridValues.Evaluate(policy, rewarder);
genericValues.Evaluate(policy, rewarder);
Assert.That(() => AllValuesAreEqual(gridWorld, genericValues, gridValues));
}
