Hybrid microgrids (HMGs) typically combine renewable-based and conventional generation with energy storage and controllable loads, offering a number of benefits besides simple reduction of CO2 emissions. If properly sized and controlled, such HMGs could be operated as fully dispatchable parts of a wider distribution network and could significantly improve overall system reliability and reduce impact of outages on the connected customers. Designing and operating HMGs, however, is a complex task and this paper presents an approach based on improved binary genetic algorithm (IBGA) method for optimal dispatching and control of all HMG resources: renewable and conventional generators, energy storage system and demand-manageable loads. Following an outage and based on available HMG resources, the IBGA aims to find a solution for network reconfiguration in which radial structure of the network is maintained and loads are supplied in accordance to a specified priority list, while also satisfying operational constraints: bus voltage and branch loading limits. The presented methodology is illustrated and validated on a commonly used IEEE 33-bus test network, where obtained results demonstrate that HMGs can improve overall system reliability and reduce negative impact of outages on the connected customers.