This research focuses on event varying discrete event systems with capacity and order constraints, and derives two state-space representations that stand for the earliest and latest times of event occurrences. Moreover, we consider rescheduling methods for these representations that can be applied even when the relevant parameters are changed after the commencement of the job. The two state-space representations derived have forms that are similar to dual systems in modern control theory. In online scheduling, it is often essential to keep track of state changes or the float times of jobs continuously. For the calculation of the float times, both earliest and latest times must be calculated. Hence, the derived representations are an attractive tool for solving online scheduling problems.