Integrating Project Recommendations from Competing Master Plans Under Limited Budgets

PURPOSE The City of Atlanta Department of Watershed Management (DWM) is moving to a 5-year rolling Capital Improvement Program (CIP) rather than annual CIP development. As part of this transition, the Atlanta Program Management Services Team (PMST) was tasked with developing the Atlanta Integrated Water Resources Plan (IWRP) to incorporate project recommendations from the City's three recently completed master plans for water, wastewater, and stormwater. This effort was especially difficult as the City's available budget for CIP projects was being significantly reduced from normal years because of the adverse revenue impacts associated with the coronavirus pandemic and the uncertain economic recovery forecasts for the 5-year rolling CIP time frame. This paper details the successful development of an optimization model designed to maximize triple bottom line (TBL) and risk reduction benefits from the universe of potential water, wastewater, and stormwater projects while meeting tight financial budget limitations. The optimization model was based on OptimaticsTM. software that uses a heuristic learning algorithm, which is an approach designed to solve multi-criteria problems in a faster and more efficient manner that favors speed of process over absolute accuracy or completeness. The model used in Atlanta was the 3-dimensional (3-D) version to accumulate as much triple bottom line per dollar (TBL/$) and risk reduction per dollar (RRB/$) as early as possible in the planning horizon while minimizing budget expenditures. METHODOLOGY The IWRP established CIP budgetary limitations based on the City's rate consultant's projections of the total encumbrance capacity between Fiscal Year (FY) 2020 and 2027. The projects included allowances for grant and loan funding from state and federal sources and for loan proceeds from commercial paper and revenue bond issuance. The encumbrance capacity was reduced for committed project expenditures for required Consent Decree projects, annual projects typically funded with CIP rather than operational funds, an emergency or unplanned expenditure allowance, annual architect/engineering (A/E) expenditures, and projects funded by a separate Security Surcharge program. Figure 1 illustrates the projected budgets for the remaining years of the 10-year optimization planning period. Since the City does not have a stormwater utility providing a dedicated source of stormwater funds, the City has allocated up to 10 percent of its Municipal Option Sales Tax (MOST) revenue for stormwater projects. This MOST allocation is projected to end in FY 7 assuming the City's creates a Stormwater Utility that generates more revenue than needed for utility operations and maintenance, at which point, more CIP funding becomes available. At this time, the MOST revenue is returned to water and wastewater project funding needs. Project benefits were assigned to each CIP project based on either a 10-point scale (for TBL benefits) or 5-point scale (for risk reduction benefits). The TBL and risk reduction criteria were developed by DWM personnel during the IWRP development and specific scoring plans were developed to facilitate consistent scoring. Project costs were developed in each of the separate master plans and updated to the IWRP cost year based on a 3 percent per year cost escalation factor. Table 1 lists the TBL criteria for water and wastewater and Table 2 lists the TBL criteria for stormwater. Table 3 lists the risk reduction criteria. Optimizer model simulations were then used to evaluate scenario plan results based on budget conformance, cumulative TBL benefit from all constituent projects, and cumulative risk reduction benefit from all constituent projects. Model inputs included: - Project data (unique identification number or code, TBL score, RRB score, estimated cost, and project duration); - Schedule constraints (required end dates [for mandatory projects only], earliest start dates, and required project prerequisites [for any project than must be performed before starting the current project]); and - Budget constraints (annual total budget limits). Years with low CIP budget availability required 'carry-over' funds to be reserved from previous years to fund projects that extended more than one year in project duration. Due to adverse revenue impacts from the coronavirus pandemic and the uncertainties associated with economic recovery, the optimization budgets were reduced during the IWRP. To accommodate the more stringent budget limitations, the number of potential CIP projects from the master plans were also reduced to include only the most critical CIP projects. Three factors were analyzed to identify critical CIP projects, including: - Regulatory Compliance considered projects critical if failure to implement the project causes the utility to violate permit conditions; - Single Point of Failure considered projects critical if the project impacts a single point of failure asset or process such that failure to implement the project causes the asset or the process to fail; and - Positive Revenue Impact considered projects critical if implementation of the project increases revenue or significantly reduces costs thereby freeing revenue for implementation of other CIP projects. For water and wastewater projects, at least two of the criticality factors had to be identified. For stormwater projects where few projects met the two-criticality factor threshold, only once criticality factor was required to be considered a critical project. RESULTS Unlike many models, the OptimizerTM model develops several potential combinations of CIP project plans, termed scenarios, for selection of the 'optimal' plan by the utility. Some plans may have higher TBL benefits but lower risk reduction. Others may have better risk reduction benefits but lower TBL scores. For the final Optimizer runs in the IWRP, the optimal plan selection had the best TBL benefit score and the best risk reduction score but left just under $10 million in unspent budget over the 10-year planning period. Other plans were available that spend more of the budget, but the TBL and risk reduction scores were lower. Thus, it was deemed better to use unspent budget as an additional contingency for possible cost impacts and unforeseen issues. Figure 2 illustrates the CIP project breakdown between water, wastewater, and stormwater projects for the first 5-year rolling CIP period. The numbers of active projects by type are shown below the bar graph expenditures. FY 1 is not included because CIP decisions were made while the IWRP was in progress. Figure 3 illustrates the full planning period breakdown. The 'negative costs' below the origin starting in FY 7 are the estimated Stormwater Utility revenues that start to fund stormwater projects, if implemented. The optimization model is designed to run on a routine and periodic basis. The CIP project components and cost assumptions must be reviewed and updated as needed. Similarly, the TBL and risk reduction project scores should be updated if project components or conditions change. As with all models, the optimization model's usefulness is dependent on the accuracy of the data input for processing. CONCLUSIONS Use of a rolling CIP helps facilitate fair and equitable evaluations by serving as a formalized process for including new CIP projects in the adopted CIP. Use of the optimization model to evaluate the costs versus TBL versus risk reduction benefit of each project allows the most beneficial and economic projects to be selected for implementation. While developing the model, DWM planners and managers realized the varying levels of accuracy and care in defining potential CIP projects varied across the organization. It was widely recognized that using a more formalized process would foster more consistency and accuracy when developing project recommendations. In a restrictive budget period when only the most critical projects can be advanced, it is even more important the DWM select the most beneficial projects that addresses the utility's best needs.