Lukas' Notes

Definition

NP-Optimisation Problem

An NP-optimisation problem is an optimisation problem

where denotes the binary encoding length of :

ComponentMeaningPolynomial-time requirement
Valid instancesDecide
Feasible solutions of Decide
Objective value of feasible Compute

Time is measured in the input length: or , respectively. For some fixed polynomial ,

Task: given with , find an optimal solution

Intuition

Candidates are short; instance validity, feasibility, and objective value are efficiently checkable. Finding the best candidate need not be efficient. Membership in NP-optimisation alone does not imply NP-hardness.

Decision Version

Definition

NP-Optimisation Decision Problem

For an NP-optimisation problem , the associated decision problem is:

Instance: with and , both binary-encoded.

Question: does a feasible solution meet the threshold?

Link to original

Comparison to NP

Both NP and NP-optimisation use polynomially balanced, polynomially decidable instance–candidate relations. NP asks whether an accepting candidate exists; NP-optimisation asks which feasible candidate is best.

NPNP-optimisation
InputInstance Instance
OutputYes / noOptimal solution , or infeasibility
Polynomial-time checkCertificate acceptanceFeasibility and objective value, not optimality
Extra structureNone requiredObjective and goal or