According to the concept of physical integration as understood in axiomatic design, design parameters of a product should be integrated into a single physical part or a few parts with the aim of reducing the information content, while still satisfying the independence of functional requirement. However, no specific method is suggested in the literature for determining the optimal degree of physical integration in a given design. This is particularly important with the current advancement in technologies such as additive manufacturing. As new manufacturing technologies allow physical elements to be integrated in new ways, new methods are needed to help designers optimize physical integration given the specific constraints and conflicts of each design. This study proposes an algorithm that uses graph partitioning to allow a designer to optimize the integration of functional requirements into a target number of parts, with the goal of minimizing the co-allocation of incompatible functional requirements in the same part. The operation and viability of the algorithm are demonstrated via two numerical examples and a practical example of designing a pencil.