In 2009, Roglin and Teng showed that the smoothed number of Pareto optimal solutions of linear multi-criteria optimization problems is polynomially bounded in the number n of variables and the maximum density O of the semi-random input model for any fixed number of objective functions. Their bound is, however, not very practical because the exponents grow exponentially in the number d+1 of objective functions. In a recent breakthrough, Moitra and O'Donnell improved this bound significantly to O(n2dOd(d+1)/2).
An "intriguingproblem", which Moitra and O'Donnell formulate intheir paper, is how much further this bound can be improved. The previous lower bounds do not exclude the possibility of a polynomial upper bound whose degree does not depend on d. In this paper we resolve this question by constructing a class of instances with Ω((nO)(d-log(d))ċ(1-Θ(1/O))) Pareto optimal solutions in expectation. For the bi-criteria case we present a higher lower bound of Ω(n2O1-Θ(1/O)), which almost matches the known upper bound of O(n2O).
[1]
D. Spielman,et al.
Smoothed analysis of algorithms: Why the simplex algorithm usually takes polynomial time
,
2004
.
[2]
Shang-Hua Teng,et al.
Smoothed analysis: an attempt to explain the behavior of algorithms in practice
,
2009,
CACM.
[3]
Shang-Hua Teng,et al.
Smoothed Analysis of Multiobjective Optimization
,
2009,
2009 50th Annual IEEE Symposium on Foundations of Computer Science.
[4]
Matthias Ehrgott,et al.
Multicriteria Optimization
,
2005
.
[5]
René Beier,et al.
Random knapsack in expected polynomial time
,
2003,
STOC '03.
[6]
René Beier,et al.
The smoothed number of Pareto-optimal solutions in bicriteria integer optimization
,
2007,
Mathematical Programming.
[7]
Ankur Moitra,et al.
Pareto optimal solutions for smoothed analysts
,
2010,
STOC '11.