[GavinMcG]

You're right that you need some measure to judge improvement, but you can have a vision of perfection without making it a goal in itself.

Optimiziation may be a subset of problem solving, but surely you'd agree that a problem can be solved even without the situation being brought to optimality?

[TeMPOraL]

Optimization is not just about reaching optimal state - which for all practical problems is impossible, given both the constraints of physical reality and the fuzzy definition of optimum humans can conceptualize. Optimization is a process of determining what the optimum is and moving the system towards that point.

[Retra]

If perfection is not a goal, how is it useful? And why call it perfection? I'd argue that you can have models without making them goals, but a model is not perfection unless it is also your primary goal. Perfection implies an optimal moral evaluation, and moral evaluation is always relative to a goal.