Robotics needs to fulfill complex requirements for real-world implementations. Such requirements include heterogeneous functionality in a swarm of robots and each robot cannot be always equipped with all necessary functions due to limited resources. How many robots involve a certain function and how many others are reserved for another function are a difficult problem. The nature/bio-inspired computing community provides robotics with solutions for many problems in the form of optimization algorithms. However, those algorithms are mainly formulated under an assumption that each computational agent is defined to be uniform or homogeneous to each other. In an article recently published in the Journal of Applied Artificial Intelligence, we analyze how a certain ant species including many immobile individuals accomplishes efficient migration and introduce a simulation algorithm wherein two groups of different types of agents representing functional heterogeneity acquire the optimal solution for target search through iterative computational process, while preserving the performance dynamics efficiency in comparison with the single group of agents performing a homogeneous function. We show that the simulation results correspond to field experiments reported by external experts of the species and those results are fully supported by both a general statistical-test and a quantitative evaluation method proposed in robotics.