# Question 6: Evaluation Functions for Two-Player Games (15 Marks)
Improve upon your Minimax agent by implementing a strong _evaluation function_
for this game. The evaluation function will be invoked when the search cannot
Copyright By PowCoder代写 加微信 powcoder
reach a terminal state within the given **depth limit**.
You need to implement your evaluation function in such a way that you can use
it within the method `evaluation()` of the class `MinimaxAgent` in
[minimax_agent.py](../minimax_agent.py). In addition, you will have to create
and include a text file called `eval_function.results.md` whose contents are
detailed below. There is no need to include your student ID in this file.
To give you an idea of what kinds of scores you should aim for, here are the
numbers achieved by our benchmark evaluation function. You don’t have to
achieve these scores to get full marks. We will manually look at your code and
give marks based on originality, simplicity, and efficiency. Thus you’ll need
to comment your code well so that the tutor can understand your method.
| Maze | Depth | Score |
| ——————— | —– | —– |
| testAdversarial | 12 | 89 |
| smallAdversarial | 2 | 92 |
| aiAdversarial | 10 | 86 |
| anuAdversarial | 8 | -78 |
| mazeAdversarial | 10 | -79 |
| smallDenseAdversarial | 6 | 253 |
| aiDenseAdversarial | 6 | -504 |
| anuDenseAdversarial | 6 | -579 |
| mazeDenseAdversarial | 6 | 702 |
We need you to include in the file `eval_function.results.md` a table like the
one above, encoded using Markdown syntax (see the source of this file to see
how to do that), reporting your depths and scores. There is no need to
automatically generate this file. You can simply manually put your results into
the table of the md file.
You can do the tests with the following command:
python3 red_bird.py -p MinimaxAgent -l adv_search_layouts/