A Swarm Production System (SPS) aims to be an agile and resilient Reconfigurable manufacturing system (RMS) paradigm that incorporates mobile workstations and transport robots on the factory production floor. This paper primarily focuses on SPS’s initial but recurring planning stage termed topology planning, which dynamically changes throughout the production runtime with spatially adaptive workstations and transporters handled exclusively by a Topology Manager (TM). TM is essential to multi-variant production with the optimal positioning of the workstations and provides a topology that optimizes the traffic flow for the product carrier robots. TM is a bridge to enable SPS to integrate with general planning and scheduling systems like ERP and MES and is comprised of a Topology Planner (TP) that evaluates the ideal configuration of on factory floor for a batch of product mix and a Reconfiguration Decision System (RDS) that decides on applying the estimated new topology during the batch changeover. The paper proposes a framework for the TM to identify its essential functionalities, responsibilities and working principle in a swarm production system. The paper also describes a grid-based heuristic approach applicable to two-dimensional spatial problems to reduce the complexity of the NP-hard problem. The paper focuses on a framework to estimate a reconfigurable shop floor layout with a Force-directed Graph-theory algorithm. A stochastic statistical model evaluates the performance of the optimal topology for throughput and makespan.