Application of Bat-Inspired Computing Algorithm and Its Variants in Search of Near-Optimal Golomb Rulers for WDM Systems: A Comparative Study

Abstract

The algorithms inspired by the nature are the powerful computing algorithms in solving various NP-complete industrial and engineering design problems. This chapter presents a comparative study of bat-inspired computing algorithm and its hybrid variants in order to discover near-optimal Golomb rulers (OGRs). Near-OGR sequences can be used as a channel allocation scheme to reduce one of the important nonlinear crosstalk generated via four-wave mixing (FWM) signals in an optical wavelength division multiplexing (WDM) systems. The OGRs provide unequally spaced channel allocation, a bandwidth-efficient scheme, then the uniformly spaced channel allocation methods to minimize the FWM crosstalk signals. To explore the search space, the bat-inspired computing algorithm is hybrid in its simple form with differential evolution (DE) mutation and random walk characteristics. The algorithms solve the two parameters, namely, the length of the Golomb ruler and total unequally spaced channel bandwidth occupied by OGRs in the optical WDM systems. The results reveal that the presented bat-inspired computing algorithm and its variants are better than other classical computing methods such as extended quadratic congruence (EQC) and search algorithm (SA) and nature-inspired computing algorithms, namely, genetic algorithms (GAs), and simple big bang–big crunch (BB–BC) computing algorithm to generate near-OGRs in terms of the length of ruler, the total occupied channel bandwidth, the bandwidth expansion factor (BEF), the CPU time and the computational complexity. This comparative study also concludes that the hybridization with both DE mutation and random walk schemes likely outperforms other methods for large mark values.