Solving the Convex Cost Integer Dual Network Flow Problem
Management Science
Weighted isotonic regression under the L1 norm
SODA '06 Proceedings of the seventeenth annual ACM-SIAM symposium on Discrete algorithm
Computational Optimization and Applications
An improved earliness-tardiness timing algorithm
Computers and Operations Research
Discrete Applied Mathematics
Unimodal regression via prefix isotonic regression
Computational Statistics & Data Analysis
Journal of Computer and System Sciences
Learning Bayesian network parameters under order constraints
International Journal of Approximate Reasoning
Nonparametric combinatorial regression for shape constrained modeling
IEEE Transactions on Signal Processing
Some new results on inverse sorting problems
COCOON'05 Proceedings of the 11th annual international conference on Computing and Combinatorics
Hi-index | 0.00 |
We consider the problem of minimizing ? j?NC j(x j), subject to the following chain constraints x 1 =x 2 = x 3 = 脗·脗·脗· = x n, where C j(x j) is a convex function of x j for each j ? N={1,2, ... ,n}. This problem is a generalization of theisotonic regression problems with complete order, an important class of problems in regression analysis that has been examined extensively in the literature. We refer to this problem as thegeneralized isotonic regression problem. In this paper, we focus on developing a fast-scaling algorithm to obtain an integer solution of the generalized isotonic regression problem. LetU denote the difference between an upper bound on an optimal value of x n and a lower bound on an optimal value of x 1. Under the assumption that the evaluation of any function C j (x j) takesO(1) time, we show that the generalized isotonic regression problem can be solved inO(n logU) time. This improves by a factor of n the previous best running time ofO(n 2 logU) to solve the same problem. In addition, when our algorithm is specialized to theisotonic median regression problem (where C j(x j)=c j|x j-a jI) for specified values ofc js anda js, the algorithm obtains a real-valued optimal solution inO(n logn) time. This time bound matches the best available time bound to solve the isotonic median regression problem, but our algorithm uses simpler data structures and may be easier to implement.