Optimization in the Loop: Implementing and Testing Scheduling Algorithms with SimuLTE

TL;DR

This paper demonstrates how to integrate commercial optimization solvers like CPLEX into OMNeT++ simulations, enabling real-time testing of scheduling algorithms in LTE network models such as SimuLTE.

Contribution

It introduces two methods for embedding CPLEX into OMNeT++, facilitating dynamic optimization within network simulations.

Findings

Successful integration of CPLEX with OMNeT++ for scheduling

Real-time testing of LTE resource scheduling algorithms

Comparison of external solver and API-based approaches

Abstract

One of the main purposes of discrete event simulators such as OMNeT++ is to test new algorithms or protocols in realistic environments. These often need to be benchmarked against optimal/theoretical results obtained by running commercial optimization solvers. The usual way to do this is to have the simulator run in a standalone mode and generate (few) snapshots, which are then fed to the optimization solvers. This allows one to compare the optimal and suboptimal solutions in the snapshots, but does not allow to assess how the system being studied would evolve over time if the optimal solution was enforced every time. This requires optimization software to run directly in the loop of the simulation, exchanging information with the latter. The goal of this tutorial is to show how to integrate a commercial solver (CPLEX) into the simulation loop of the OMNeT++ environment. For this purpose, we propose two methods: a first one that uses a solver as an external program, and a second one that exploits a C-written API for CPLEX known as Callable Library. We then exemplify how to apply these two methods to SimuLTE, a simulation model for LTE cellular networks, implementing and testing a simple solution to a well-known resource-scheduling problem.