Suggestions for more stop criterias

[caranha]

Some simple to program stop criterias, for adding after the paper is submitted:

• diversity based stop criteria (number of identical solutions, or distance between new and old population)
• utility based stop criteria (utility threshold of stopping the run, or delta utility between previous iteration and current one)

[fcampelo]

Your "diversity-based" actually describes two different criteria: diversity-based and improvement-based.

For MOBJ I don't think population diversity will be a good stop criterion, except in very particular situations (problems with non-tradeoff objectives).

Improvement based could be a good idea, and is related to your other suggestion (utility-based) - we would only need to define a utility measure (hypervolume, maybe? Or another indicator?)