Case studies

Business impact

The Fire Program Analysis (FPA) project leverages IBM expertise in advanced analytics to help improve management of federal wildlands while meeting multiple objectives. The five agencies involved are using a sophisticated tool from IBM Research to analyze wildland fire management priorities and requirements, and improve strategic investment decision making.

Issue

Five federal agencies are involved in managing federal wildlands: the U.S. Forest Service; the Bureau of Indian Affairs; the Bureau of Land Management; the U.S. Fish and Wildlife Service; and the National Park Service. Because the agencies have a mix of land management goals and associated budgetary requirements, FPA was federally mandated by Congress to help coordinate the agencies’ program budget development.

Executive summary

IBM Research Services—a partnership between Global Business Services (GBS) and IBM Research—developed a strategic budgeting tool based on multi-objective optimization-based modeling technology to help analyze and evaluate the future requirements of the five agencies. The project’s initial phase—begun in 2003—focused on predicting budgetary requirements for the agencies; since then the scope has expanded to using the modeling technology to develop a more holistic view of the agencies’ strategic objectives and associated investment requirements.

What IBM did

Through IBM Research Services, GBS has worked with IBM Research to develop the FPA strategic budgeting tool capable of evaluating multiple goals and funding alternatives to help best achieve overall objectives. The simultaneous analysis of multiple and sometimes conflicting objectives requires the use of the multi-objective LP (linear programming) modeling and optimization approach rather than traditional linear programming or mixed-integer programming methods.

As part of the GBS solution, experts at IBM Research developed a multi-objective LP model to analyze how different levels of investment in a fire planning unit may increase or decrease, and how well the mix of resources (engines, crews, bulldozers, helicopters, etc.) is able to meet FPA performance measures. The FPA tool can also analyze the investment decisions for reactive preparedness (managing fires) vs. a proactive approach (managing vegetation to prevent future fires) for all five agencies. The resulting information allows the agencies to analyze the impact of their preferences (budget constraints and multiple-goal targets) to compute the optimal investment alternatives that best achieve both near- and long-term goals.

In addition to project management, the GBS team developed and implemented the IT infrastructure architecture and system design necessary for the multi-objective LP modeling and analysis. The team also oversaw the agencies’ data gathering, storage and transformation activities. With help from GBS consultants, the agencies were able to understand the impact and costs of activities such as fire suppression, reduction of wildland-urban interface acreage burned, protection of highly valued resources and increases in the initial attack success rate.

The FPA tool can be adapted to accommodate additional factors for analysis. Future possibilities could include enhancing the model to analyze an expanded set of performance measures, including carbon management, or using the tool to analyze investment decisions for states and geographic regions for use in localized analysis.

Capabilities applied

The combination of IBM Research’s capabilities in multi-objective LP modeling and GBS’s experience in enterprise architecture and application development have resulted in a flexible tool for complex analysis of multiple and conflicting objectives involving multiple factors. The tool’s flexible hybrid goal programming algorithm also makes it useful for broader use in strategic budgeting scenarios.