Multi-Level Grammar Genetic Programming for Scheduling in Heterogeneous Networks Takfarinas Saber 1( ) , David Fagan 1 , David Lynch 1 , Stepan Kucera 2 , Holger Claussen 2 , and Michael O’Neill 1 1 Natural Computing Research and Applications Group, School of Business, University College Dublin, Ireland {takfarinas.saber, david.fagan, m.oneill}@ucd.ie, david.lynch@ucdconnect.ie 2 Bell Laboratories, Nokia, Dublin, Ireland {stepan.kucera, holger.claussen}@nokia-bell-labs.com Abstract. Co-ordination of Inter-Cell Interference through scheduling enables telecommunication companies to better exploit their Heteroge- neous Networks. However, it requires from these entities to implement an effective scheduling algorithm. The state-of-the-art for the scheduling in Heterogeneous Networks is a Grammar-Guided Genetic Programming algorithm which evolves, from a given grammar, an expression that maps to the scheduling of transmissions. We evaluate in our work the possi- bility of improving the results obtained by the state-of-the-art using a layered grammar approach. We show that starting with a small restricted grammar and introducing the full functionality after 10 generations out- performs the state-of-the-art, even when varying the algorithm used to generate the initial population and the maximum initial tree depth. Keywords: Telecommunication, Heterogeneous Network, Scheduling, Grammar- Guided Genetic Programming, Multi-Level Grammar. 1 Introduction We have seen in the last decade a proliferation in the use of mobile phones worldwide to reach 4.47 billion users in 2017 and this number is expected to exceed the 5 billion barrier by 2019 [1]. Companies attempt to attract new costumers through price cuts and the introduction of new technologies, like the soon-to-come 5G networks. Due to the heterogeneity and growing size of the networks, there is a large and increasing need to optimise their performance [2]. In traditional single cellular networks, Macro Cells (MCs) are employed to cover all User Equipments (UEs) such as phones, tablets, and any other device equipped with a broadband adapter. However, with the explosion of connected devices, MCs struggle to cope with the load. Therefore, they have to be sup- plemented with local and less powerful Small Cells (SCs), creating a two-tiered configuration called Heterogeneous Networks (HetNets). SCs are deployed in ar- eas with traffic hot-spots to attract the near-by UEs, which offloads the MCs and