The purpose of this study is to examine the effect of individual components' reliability on a system's reliability. The system refers to the Financial Management Information System (FMIS), the US Federal Highway Administration's (FHWA) web‐based project approval and tracking software. Its components are 61 project information fields.
The analysis would view each highway project‐funding request as an activity with cycle‐dependent performance for which success probability can be calculated as Reliability, R. The reliability analysis of the 61 FMIS fields results in a series system with Rsys the “estimated reliability” of finding “true” values in all 61 information fields during one highway‐related project funding authorization review.
Of an estimated 200 projects approved, there was previously estimated a 50 percent to 80 percent unreliability rate, while the study found an unreliability rate of approximately 80 percent.
Owing to the nature of federal government software, data can be very difficult to acquire in this working environment, but a simple calculation was relatively successful in confirming the “estimated reliability” of finding “true” values and showing how the reliability could dramatically decrease.
The paper contributes to the applicability of reliability analysis to project approval software, showing the progression from estimated data to bounding the estimate using reliability theory.
