Current methods of optimal power flow were not designed to handle increasing level of volatility in the electrical networks, this thesis suggests that a message passing-based approach could be useful for managing power distribution in electricity networks. This thesis shows the adaptability of message passing algorithms and demonstrates and validates its capabilities in addressing scenarios with inherent fluctuations, in minimising load shedding and generation costs, and in limiting voltages. Results are promising but more work is needed for this to be practical to real networks.