Routing, Modulation Level, Spectrum and Multiwavelength Transponder Assignment in IP-over-EON – Integer Linear Programming and Genetic Algorithm Approaches

Abstract

Elastic Optical Networks (EONs) represent a high-spectral-efficiency paradigm for accommodating heterogeneous traffic demands. However, EONs impose stringent physicallayer operational constraints, including spectrum non-overlapping, contiguity, and continuity, collectively designated as the Routing and Spectrum Assignment (RSA) problem. The integration of adaptive modulation schemes enhances spectral utilization by dynamically allocating resources according to transmission distance and demand characteristics, thereby extending the problem formulation to Routing, Modulation Level, and Spectrum Assignment (RMSA).In IP-over-EON architectures, the aggregation and grooming of low-bandwidth IP traffic flows are critical for optimal utilization of optical channel capacity. The deploymentof Multi-wavelength Sliceable Bandwidth Variable Transponders (MW-SBVTs), which enable simultaneous multi-carrier transmission within bounded spectral tuning windows,introduces additional complexity. This necessitates explicit joint transponder allocation, giving rise to the Routing, Modulation Level, Spectrum, and Multi-wavelength Transponder Assignment (RMLSTA) problem.This work addresses the static RMLSTA problem in IP-over-EON architectures through two complementary methodologies: a sequential multi-phase Integer Linear Programming (ILP) formulation and an efficient Multi-Objective Genetic Algorithm (GA). Given an IP traffic demand matrix and physical network topology, both approaches optimize the resource allocation following a lexicographic multi-objective optimization framework: minimizing the maximum utilized spectrum index, the total number of allocated frequency slots, and the number of active transponders. The proposed multi-phase ILP formulation decomposes the intractable joint problem into four consecutive optimization stages, solving each subproblem to exact stage-wise mathematical optimality. Extensive numerical simulations demonstrate that the multi-phase ILP achieves superior spectral and transponder utilization at the expense of computational scalability under high traffic loads. Conversely, the proposed GA metaheuristic achieves a favorable and robust trade-off between solution quality and computational efficiency, generating highly competitive resource assignments within predictable execution times.