A Criterion Space Search Algorithm for Biobjective Mixed Integer Programming: the Boxed Line Method
Tyler Perini(perinitagatech.edu )
Abstract: Criterion space search algorithms for multiobjective integer and mixed integer programming have gained in popularity in the past decade and many of the fastest algorithms for generating the nondominated frontier belong to this class of algorithms. We propose a new criterion space search algorithm for solving biobjective mixed integer programs: the Boxed Line Method. For one variant of the algorithm, we show that the number of single-objective mixed integer programs solved is a linear function of the number of line segments in the nondominated frontier. A computational study demonstrates that the algorithm also performs well in practice, outperforming existing algorithms.
Keywords: biobjective mixed integer programming, criterion space search, performance analysis
Category 1: Applications -- OR and Management Sciences
Citation: H. Milton Stewart School of Industrial and Systems Engineering Georgia Institute of Technology, Atlanta, 09/2017.
Entry Submitted: 09/23/2017
Modify/Update this entry
|Visitors||Authors||More about us||Links|
Search, Browse the Repository
Give us feedback
|Optimization Journals, Sites, Societies|