Skip to content

JA8. Learning Journal 8

The Learning Journal is a tool for self-reflection on the learning process. The Learning Journal will be assessed by your instructor as part of your Final Grade.

Answer the following questions

1. Describe what you did. You need to describe what you did and how you did it

This was the final week of this course; it was all about how to environments with multiple interacting agents. I started by reading the materials assigned in the learning guide, then I did the discussion assignment which was about discussing the minimax algorithm and its alpha-beta pruning technique. I also did the self quiz along with the learning journal.

2. Describe your reactions to what you did

I found the topics presented so far to be of great importance theoretically and practically, but they seem complex and hard to understand. The idea of multi-agent systems and game theory is very big and needs more time to be understood but we only touched the surface of it. The discussion assignment mentioned the minimax algorithm and its alpha-beta pruning technique, but I think we should have a programming assignment to implement them and practice them in a real-world scenario.

3. Describe any feedback you received or any specific interactions you had while participating discussion forum or the assignment Discuss how they were helpful

I think that this week is putting everything we learned so far in this course together in one place. The discussion assignment ensured that by asking us to provide an example of the minimax algorithm and its alpha-beta pruning technique. Creating the game tree reminded me of graphs, states, actions, searching graphs, features and constrain functions, knowledge representation, reasoning, planning, and decision-making. I think I was happy because I have to touch each of these topics to conclude my example.

4. Describe your feelings and attitudes

I was surprised that the game representation methods come from economics, especially the normal and extensive form games. It is a good example of how different fields can intersect and inform each other. I also found it interesting that the minimax algorithm is based on the assumption that the opponent will always play optimally. This is a strong assumption, and in real-world scenarios, opponents may not be familiar with the game or may make mistakes that deviate from optimal play.

The alpha-beta pruning technique is a clever optimization that can significantly reduce the number of nodes evaluated in the game tree, making the minimax algorithm more efficient. Although the degree of pruning is typically unknown in advance, this example of conditional probability pruning shows good results in practice.

5. Describe what you learned. You can think of one or more topics and explain your understanding in writings

Multi-agent systems are systems in which multiple agents interact with each other to achieve their goals. Those agents can be cooperative or competitive, and they can have different levels of knowledge about the environment and each other. Cooperative agents have common goal(s) and they share information and coordinate their actions to achieve mutual benefits. Competitive agents, on the other hand, have conflicting goals and they may try to outsmart or outmaneuver each other to maximize their individual utility, which usually happens at the expense of other agents (Poole & Mackworth, 2017).

There are three representations of games: normal form games, extensive form games, and multi-agent decision networks. The normal form game is a matrix representation of the game where each cell represents the utility values for each player given their actions. The extensive form game is a tree representation of the game where each node represents a game state and each edge represents a possible move. The multi-agent decision network is a directed acyclic graph (DAG) that includes potential states and the probability (belief) that this state will occur along with the dependencies between these states (Poole & Mackworth, 2017).

The mini-max algorithm is a decision-making algorithm used in game theory in 2-player zero-sum games. In these games, there are two agents (players) competing against each other, and their will be only one winner, that is, one player’s win means the other’s loss. The purpose of the algorithm is to find the optimal move for a player, assuming that the opponent will also play optimally.

6. Did you face any challenges while doing the discussion or the development assignment? Were you able to solve it by yourself?

The discussion assignment asked us to explain the minimax algorithm and its alpha-beta pruning technique providing an example of how it is being used. I used XO game (tic-tac-toe) as an example where I manually created a game tree and explained one round of running the algorithm to choose 1 move, I’m attaching the graph here:

Mini-Max Algorithm

The challenge I faced was coming up with the right example so I chose a simple game. The next challenge was drawing the graph which was super complex due to the many branches and nodes where ech of them represent the entire game. I used Mermaid language to draw the graph in a Markdown file which made it easier to visualize the game tree.

References