Field-deployed robotic fleets can provide solutions that improve operational efficiency, control operational costs, and provide farmers with transparency over the day-to-day operations with scouting operations. The topology of agricultural farms such as polytunnels provides a basic environmental configuration that can be exploited to create a topological map to aid operational planning and robot navigation. However, these environments are optimised for operations by humans or for large farming vehicles and pose a major challenge for multiple moving robots to coordinate their navigation while performing tasks. The farm environment without any topological modifications for supporting robotic fleet deployments can cause traffic bottlenecks, eventually affecting the overall efficiency of the fleet. In this work, we propose a Genetic Algorithm-based Topological Optimisation (GATO) algorithm that discretises the search space of topological modifications into finite integer combinations. Each solution is encoded as an integer vector that contains the location information of the topology modification. The algorithm is evaluated in a discrete event simulation of the picking and in-field logistics process in a commercial strawberry farm and the results validate the effectiveness of our algorithm in identifying the topological modifications that improve the efficiency of the robotic fleet operations.