While autonomous guided vehicle systems are increasingly used in homogeneous and structured environments, their use in complex and variable scenarios is usually limited. Established algorithms for the navigation of systems use static maps with deterministic metrics, which can only achieve optimal results in clearly defined environments. In dynamic and extensive deployment scenarios, which are also dependent on a large number of influencing parameters, autonomous intralogistics systems cannot yet be deployed dynamically. One example here is mixed transport between buildings under changing weather conditions. As a solution for dynamic navigation, we propose a hybrid metric in combination with topological maps and cyclic environmental sensing. Based on a quantification of influencing factors on each intralogistics entity, an optimal and dynamic navigation of every system can be performed at any time. The individual components are implemented in the context of an autonomous tow truck system and evaluated in different application scenarios. The results show significant added value in use cases with sudden weather changes and complex route networks.