PPSN XII - LNCS 7491-7492 CD-ROM

A Benchmark Generator for Dynamic Permutation-Encoded Problems

Michalis Mavrovouniotis¹, Shengxiang Yang², and Xin Yao³

¹Department of Computer Science, University of Leicester, University Road, Leicester LE1 7RH, United Kingdom
mm251@mcs.le.ac.uk

²Department of Information Systems and Computing, Brunel University, Uxbridge, Middlesex UB8 3PH, United Kingdom
shengxiang.yang@brunel.ac.uk

³CERCIA, School of Computer Science, University of Birmingham, Birmingham B15 2TT, UK
x.yao@bham.cs.ac.uk

Abstract. Several general benchmark generators (BGs) are available for the dynamic continuous optimization domain, in which generators use functions with adjustable parameters to simulate shifting landscapes. In the combinatorial domain the work is still on early stages. Many attempts of dynamic BGs are limited to the range of algorithms and combinatorial optimization problems (COPs) they are compatible with, and usually the optimum is not known during the dynamic changes of the environment. In this paper, we propose a BG that can address the aforementioned limitations of existing BGs. The proposed generator allows full control over some important aspects of the dynamics, in which several test environments with different properties can be generated where the optimum is known, without re-optimization.

LNCS 7492, p. 508 ff.

Full article in PDF | BibTeX