2011-MM20

Posted by

A Methodology to Improve the Predictability of the CER with Insufficient Data in Korean Weapon System R&D Environment

Models and Methods Track

MM20A_Paper_ImprovePredictabilityofCERwithInsufficientDatainKoreanWeaponSystems_Kim

MM20_Presentation_ImprovePredictabilityofCERwithInsufficientDatainKoreanWeaponsSystems_Kim

Abstract:

Parametric cost estimating models have been used widely to obtain appropriate cost estimates in early weapon system acquisition process. Parametric cost estimating models are composed of some CER(Cost Estimating Relationship)s based on regression analysis with historical data.

However, there are many restrictions in developing Korean version CER because of the insufficient number of projects and the abnormal data characteristic such as multicollinearity, outlier existing, heteroscedasticity, etc. So, the diverse regression methods have been studied in Korea to improve the predictability and stability of the each CER respectively.

We proposed a CER development process suitable for Korean
weapon system R&D environment and developed a new combining method of each regression model which is able to provide better predictive ability of CER. A real world data from historic weapon system R&D records is used to verify the performance of the developed method. Our linear combining CER method was more accurate than each regression model.

Our study will provide an appropriate methodology to develop our CERs and a more predictive method in Korean weapon system R&D environment.

Author(s):

Dong Kyu Kim
Korea National Defense University
Dong Kyu Kim is to line of duty in Republic of Korea Army for 11 years since graduate the Korea Military Academy. As he works for 3 years as an army finance officer, he accumulated business experience connected with cost and expense, and graduated officer’s basic/advanced course.

During he works in DAPA(Defense Acquisition Program Administration) for 3 years, he worked in defense acquisition policy, especially M&S(Modeling & Simulation) and achieved various studies about program management, cost-effectiveness analysis and proposed main policies about this field.

At the present, he is studying doctor’s course of operational analysis at KNDU(Korea National Defense University). Specially, he is interested in cost analysis because of experience of military program management and cost.

Now, he is studying cost analysis methodology suitable for Korean defense acquisition environment.

Sung Jin Kang

Yong Bok Lee