Linear Programming with Inexact Data is NP‐Hard
read more
Citations
Linear Optimization Problems with Inexact Data
Weak and strong solvability of interval linear systems of equations and inequalities
Nonnegative matrix factorization : complexity, algorithms and applications
Solvability of systems of interval linear equations and inequalities
Interval systems of max-separable linear equations
References
Compatibility of approximate solution of linear equations with given error bounds for coefficients and right-hand sides
Checking robust nonsingularity is NP-hard
Strong solvability of interval linear programming problems
Related Papers (5)
Frequently Asked Questions (6)
Q2. what is the problem of checking regularity of interval matrices?
But since the problem of checking regularity of interval matrices is NP hard Poljak and Rohn Theorem the problem of checking whether each system with data satisfying has a solution is NP hard as well For an m n interval matrix AI and an intervalm vector bI consider the family of LP problemsminfcTx
Q3. what is the NP hardness of the problem?
Thus the NP hardness of their problem has nothing to do with the amount of uncertainty in the data it is caused by the exponential number of vertices of an interval matrix AI Poljak and RohnNevertheless the worst case type result of Theorem does not preclude e cient solvability of many practical examples
Q4. what is the problem of existence of optimal solutions of linear programming problems in the family?
The problem of existence of optimal solutions of all linear programming problems in the family was addressed in There it was proved that each LP problem with data satisfying has an optimal solution if and only if the LP problemminfcTx Ax b Ax b x ghas an optimal solution and each of the m systems
Q5. what is the solution to the problem?
Ax b x gfor A AI b bI c e eSince the objective eTx is bounded from below a problem has an optimal solution if and only if it is feasible Hence each systemAx b xwith data satisfying A AI b bIhas a solution if and only if each LP problem with data has an optimal solution
Q6. how can the authors check regularity of AI?
In fact according to part Eq some system with data does not have a solution if and only if there exists a vector y satisfyingA c A cyjyjand eT jyjwhich is equivalent to jA cyjjyjand yThen the Oettli Prager theorem see the reformulation in Lemma gives that is equivalent to existence of a singular matrix in AIA cA cThis proves the assertion Given a square m m interval matrix AI construct an m m interval matrix AI and an interval vector bI by This can be done in polynomial time According to part checking regularity of AI can be reduced in polynomial time to checking solvability of all systems