This function returns the index (greater than or equal to 0 (zero)) of the unbound constrained variable with the minimal difference between the maximal possible value and the next maximal possible value from among the constrained variables in the array of constrained variables, vars. If all those variables have been bound already, then this function returns -1.

This function is based on the principle of regret. Regret is the difference between what would have been the best possible decision in a scenario and what was the actual decision.

This function is a predefined choice function, one of the criteria that you can use to set parameters for the order in which Solver binds constrained variables during the search for a solution.