This paper presents a decentralized, mode-adaptive (DMA) guidance law for a Hamiltonian-based controller of an N-source, dc microgrid. Droop control is commonly used for decentralized control of microgrids. Unfortunately, droop control lacks the ability to autonomously adapt to the addition or removal of a source without the augmentation of an outer loop controller. Centralized control methods provide solutions to these limitations, as well as providing the ability to globally optimize the system. To their detriment, centralized control methods are not scalable, and require system-wide information to be funneled through a central controller yielding a single point of failure. The DMA power apportionment scheme presented here aims to reduce the gap between droop control and centralized control by providing a method that can operate autonomously in the event of source and bus load fluctuations as well as reduce communication requirements of centralized control, and thus increase resiliency and scalability. After development of the DMA, its performance is compared to the centrally-controlled optimal exergy destruction (OXD) power apportionment strategy, as well as a Hamiltonianbased droop control strategy. Lastly, it is shown that for a sufficiently large number of converters supplying a microgrid, the presented decentralized, mode-adaptive strategy provides an efficient and practical alternative to both droop control and centralized control schemes.Index Terms-Dc power systems, decentralized control, distributed control, microgrids, power system control 0885-8977 (c)