The manufacturing industry cannot stay competitive and survive in today's market without agile adaptation to rapidly changing customers' demands. This in turn requires redesign and re-planning of manufacturing systems more frequently and within shorter lead-time. In our previous work, a generic reconfigurable manufacturing system model was proposed, and corresponding algorithms were implemented. However, machine breakdowns and planned and unplanned maintenance were not considered. By addressing different conditions of machines, this paper uses automata theory and queuing network approaches to derive the throughput of machines that enhances the proposed virtual production line (VPL) design methodology, in which the throughput was originally computed using the deterministic machine processing time. A priority is assigned to each idle machine according to its past performance and adaptive algorithms for reconfiguration are proposed.