Implementing stop early behavior for DiscreteOpt

[eko-wibowo]

Addressing #38 I use this in A2 so the algorithms will stop after it reaches certain value for the cases that the target optimal value is known

[eko-wibowo]

@nkapila6 do you think we need something like this for the ContinuousOpt as well? I probably can modify the base class if that's the case. Thoughts?

[nkapila6]

Hi @eko-wibowo, sorry I have been busy and haven't reviewed this.

At first glance, can you share how your can_stop attribute will be used? Because iirc the opt problems have their own can_stop method which will overwrite your can_stop method here?

Maybe in your implementation, you're using this DiscreteOpt class directly?

--

For your second question: Yes, would you like to implement this for continuous as well? Like NQueens, TSP, etc.

For the future PR:
Ideally, EarlyStopping should be something along the lines of:

• User fixes number of iterations to look for
• If score does not increase in those fixed iters
• Program should terminate
  This is how it's implemented in sklearn, torch, etc.

[nkapila6]

Ahh nevermind, I see now. It's only FlipFlop that has a separate can_stop, rest all problems are dictated through the Fitness Functions.
Haven't seen the codebase in awhile now 😅

Can you add the same for FlipFlop too? I will merge it in once it's updated there as well.

Feel free to make a second PR for the continuous one!