In this study an isolated microgrid comprising both controllable and uncontrollable sources, like solar, wind, diesel generator, fuel cell, aqua-electrolyser, hydrogen storage and battery is considered. To establish an efficient resource management strategy, a central controller takes the decisions based on the status of the loads and sources. The status is obtained with the help of multi-agent concept (treating each load and source as an agent) through internet using User Datagram Protocol/Internet Protocol (UDP/IP). The decisions are transmitted to the controllable sources to regulate their power output for damping of frequency excursion following a disturbance. A control strategy is adopted to regulate the power output from the battery only during transient, resulting in a floating battery scheme in steady state. This will reduce the ampere hour rating of the battery and can improve the damping of frequency excursion following each load disturbance. In a microgrid with generation rate constraint (GRC), tuning of controller parameters and frequency bias is a nonlinear optimisation problem. Hence, this study attempts to tune the controller parameters using an evolutionary technique named bacterial foraging optimisation (BFO). The tuned gains obtained utilising BFO method give satisfactory frequency excursion following a disturbance in the microgrid.