Integrating water in process plants require the clever arrangement of wastewater reuse and the optimal pollutant removal at various stages, not only using end-of-pipe treatment as it is customary now in industry. Such systems usually reduce the intake of water, but add constraints to the treatment. However, since all these solutions usually reduce freshwater intake, they lead to savings. Thus, most of the existing methods focus on minimizing the freshwater intake assuming that the corresponding treatment cost is proportional to the flowrate and ignoring the capital involved in grassroots or retrofit cases. We have already shown (PanAmerican Chemical Engineering Conference, Buenos Aires, October 2006, and AIChE Annual conference, 2006) that a methodology that takes the project profitability into account using ROI and NPV as measures exhibits results that are different from those obtained minimizing freshwater intake. In this paper, we extend those results to the use of decentralized regeneration using more detailed regeneration models maximizing ROI over a range of maximum allowed fixed capital investments. To compare, NPV is also maximized (They do not give always the same answer) We will also look at the impact of uncertainty of process parameters and costs in this decision making process, also considering financial risk.