Quick
Background
To understand what's going on underneath the hood here's a high-level explanation of both Mini-max and Expectimax.
Minimax is a recursive backtracking algorithm that fares exceptionally well when deciding the future paths one can take. As the game progresses down the decision tree, at each step, the best possible path is calculated, in accordance with the goal you want to maximize. (In this case food).
AlphaBeta works very similarly to Minimax except that certain parts of the recursive tree are pruned along the way (AB tree). Reducing computational time and speeding up the prediction process
Fig 1. A Decision Tree Heatmap drawn out by the AB algorithm before PacStart