Possible issue in search.ipynb

[gobber]

Hello everyone.

First, I would like to say thank you for AIMA, it is really a good library. However, I guess there is an issue in search.ipynb. In TSP problem example, the condition in hill_climbing should not it be:

if problem.value(neighbor.state) >= problem.value(current.state):

instead of:

if problem.value(neighbor.state) <= problem.value(current.state):

Because, the TSP_problem.value() method is negative (inverted path cost) and you use argmax, then I conclude that you should consider a best neighbor one that is greater then the current. Am I wrong? Thanks.

[antmarakis]

Hi, good catch! In the notebook, we simply copied the code from the repository. In hill-climbing we go from best state to best state, so currently it is correct.

The solution would be to invert the path costs so that the algorithm would work well on the problem, without changing the implementation itself.