Module optimize

Returns:

a list of numbers, starting with lo, and increasing by stepsize each time, until hi is equaled or exceeded.

lo must be less than hi; stepsize must be greater than 0.

Parameters:

• f - a function that takes a single argument of some type x and returns a value of some type y
• stuff - a list of elements of type x
• comp - a function that takes two arguments of type y and returns a Boolean; it is intended to return True if the first argument is 'better' than the second.

Returns:

a pair (bestVal, bestArg), where bestArg is the element of stuff such that f(bestArg) is better, according to comp than f applied to any other element of stuff, and bestVal is f(bestArg).

The types x and y are not actual types; they're just intended to show that the types of the functions have to match up in the right way.

For example, get the team with the highest score, you might do something like

argopt(seasonScore, ['ravens', 'crows', 'buzzards'], operator.gt)

where seasonScore is a function that takes the name of a team and returns a numerical score.

Parameters:

• objective - a function that takes a single number as an argument and returns a value
• compare - a function from two values (of the type returned by objective) to a Boolean; should return True if we like the first argument better.

Returns:

a pair, (objective(x), x). x one of the numeric values achieved by starting at xmin and taking numXsteps equal-sized steps up to xmax; the particular value of x returned is the one for which objective(x) is best, according to the compare operator.

Like optOverLine, but objective is now a function from two numerical values, one chosen from the x range and one chosen from the y range. It returns (objective(x, y), (x, y)) for the optimizing pair (x,y).