Throughput Maximization in Multi-Band Optical Networks with Column Generation

TL;DR

This work tackles throughput maximization in multi-band optical networks by solving Route–Wavelength–Band Assignment (RWBA) under band-heterogeneous margins and distance-adaptive modulation. It introduces a column-generation (CG) framework that decomposes RWBA into restricted master problems and pricing subproblems, replacing exponential path-based variables with compact wavelength-configuration columns. The approach is shown to be scalable and near-optimal compared to an ILP, achieving tens of seconds of computation up to 1200 wavelengths per link and delivering substantial throughput gains when allowing band-aware modulation (RWBA) versus fixed-band RWA. Key results on DT network topologies demonstrate throughput increases of 67% (RWA) and 125% (RWBA) under appropriate conditions, highlighting the practical impact of combining CG with flexible transceivers in MB networks.

Abstract

Multi-band transmission is a promising technical direction for spectrum and capacity expansion of existing optical networks. Due to the increase in the number of usable wavelengths in multi-band optical networks, the complexity of resource allocation problems becomes a major concern. Moreover, the transmission performance, spectrum width, and cost constraint across optical bands may be heterogeneous. Assuming a worst-case transmission margin in U, L, and C-bands, this paper investigates the problem of throughput maximization in multi-band optical networks, including the optimization of route, wavelength, and band assignment. We propose a low-complexity decomposition approach based on Column Generation (CG) to address the scalability issue faced by traditional methodologies. We numerically compare the results obtained by our CG-based approach to an integer linear programming model, confirming the near-optimal network throughput. Our results also demonstrate the scalability of the CG-based approach when the number of wavelengths increases, with the computation time in the magnitude order of 10 s for cases varying from 75 to 1200 wavelength channels per link in a 14-node network. Code of this publication is available at github.com/cchen000/CG-Multi-Band.

Figures (4)

• Figure 1: Illustration of the SNR on different channel frequencies $f_w$ after the first span assuming the ISRS (blue) and no ISRS (red), respectively. The transmission margins $M_U$ and $M_L$ are obtained by differentiating the worst-case SNR for that optical band (solid gray line, RWBA) and for all bands (dashed gray line, RWA). This paper computes the upper bound of SNRs on different optical bands by simply considering two states of ISRSs.
• Figure 2: Illustration of a 4-node network with the span length on each fiber and with a table on the right illustrating the transmission capacities for 9.0 paths on 8.0 wavelengths, $w\in\{1,...,8\}$, or on 2.0 bands, $b\in\{1,2\}$. (a), (b), and (c) show the respective assignment based on path formulation and wavelength configuration formulations, $c\in\{1,2,3,4\}$.
• Figure 3: Computation time (top row) and network throughput (bottom row) for ILP, CG, FF-$k$SP, and $k$SP-FF algorithms, for DT9 and DT14 networks, and for variable numbers of wavelength channels in 15.
• Figure 4: Maximum throughput of the DT9 using a 25 GBaud transceiver with variable modulation formats compared for the RWA (red) and RWBA (blue), and for the unlimited (filled circles) and limited transceivers (unfilled squares).