Current Research on Schedulers for Aerospace Industry Software 901014
The U.S. Army Aviation Systems Command is increasingly aware of the need to generate accurate cost/schedule estimates for mission-critical software. The estimates must be done early in the life cycle, in a disciplined manner, to achieve improved cost effectiveness. A study was contracted to reconstruct existing software data for AVSCOM cost analysis. The contract effort delivered descriptive and quantified data from case studies of actual developmental software acquisitions for DOD avionics systems. This paper exploits that data in both quantitative and statistical terms.
Four formal cost estimating models were employed in this evaluation; SECOMO (Software Engineering Cost Model); SASET (Software Architecture Sizing Estimating Tool); PRICE-S (Parametric Review of Information for Costing and Evaluation ); and REVIC (Ray's Enhanced Version of Intermediate COCOMO). Estimates were then compared to the “actual” calendar months reported in each evaluated study. This paper evaluates the scheduling veracity of the models.