Budget Reconciliation Through Dynamic Programming

Abstract

Budgeting systems for U.S. Army brigade lack transparency. Total daily commits (orders) and obligations (account withdrawals) are logged in spreadsheets that are reviewed by brigade comptrollers, but time delays occur between commits and their corresponding obligations that are not tracked in these spreadsheets. Further complications arise because credits for returned equipment are not accurately identified. It can be difficult for brigade comptrollers to accurately reconcile accounts at the end of a fiscal year, and discrepancies can cause overspending or frozen assets. In this article we derive and implement an algorithm that takes a record of daily commits and obligations over a period of time and utilizes dynamic programming to identify the most likely matching between the two. The algorithm can also estimate the probability distribution of commit-to-obligation delays, thus making it a useful prediction tool. The algorithm can be adapted to a wide range of scenarios. We verify the systems performance through a series of simulations.