A water network (WN) for a batch process could be seen as a dynamic structure which changes its topology at fixed time intervals delimited by events. During each time interval, the WN could be abstracted into an oriented graph, using a secondary ranking criterion, apart from schedule. The optimization strategy implies ordering the batch units twofold: by time, according to the predefined task schedule of each water-using unit (WU), and by maximum allowable outlet concentration of contaminants, the latter being consistent with the principle of driving force equipartition across the process. This paper focuses upon two aspects: (a) the optimization of the WN topology particular to each period, in close correlation to the storage tank behavior and (b) the dynamics of this optimized topology, as the batch process runs sequentially toward its completion. The objective is to minimize the fresh water consumption by raising the internal wastewater reuse between water-using units or from the storage tank (ST). The aforementioned optimization of the mathematical model, consisting of a system of differential algebraic equations with restrictions, is carried out using genetic algorithms, which generate the internal flows appropriate for each time window structure. A synthetic network of six WUs, three contaminants, one fresh water source, and one ST is presented and optimized under various scenarios, and the obtained results are analyzed and discussed.