Made AI stronger by working on Minimax/AB pruning

[Suraj1127]

Made AI stronger by implementing and modifying the following things.

• Integrated closed spaces(ghost goats) into the evaluation function.
• Set the minimax initial variable(value) to the different value than the value returned when win condition is reached. It was creating an error of not making any move when no move is found better.

[Suraj1127]

• Previously, I had omitted movable tigers parameter from evaluation function but AI was defeated when I played with trapping tigers strategy. Then, I integrated it into the evaluation function again.
• Problem which I thought was created by that movable tiger thing was actually because of the initial value(value) thing. So, again I included and the AI became better.

[code-geek]

Why replace INF?

[Suraj1127]

To make the value variable as small as we can. Previously, it was equal to the value of the win/loss state. Ideally, it should be +/- infinity and the win/loss state should be some finite large number. We have made the win/loss state some finite large number(though written as INF). So, we need to make the value variable smaller/larger than the INF variable.

…

On Mon, 24 Sep 2018 at 22:21, Ashish Acharya ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In Engine.py <#2 (comment)>: > @@ -37,7 +38,7 @@ def minmax(self, is_max=True, depth=0, alpha=-INF, beta=INF): # find the minimum attainable value for the minimizer if not is_max: - value = self.INF Why replace INF? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2 (review)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ATrAN4jyKBrMhYS9QTziqN1VcLO_CHgCks5ueQolgaJpZM4W2uOG> .