• 한국수학교육학회지시리즈B:순수및응용수학
• /
• 제11권1호
• /
• pp.89-96
• /
• 2004

We prove de Leeuw's restriction theorem result Jodeit, Jr. [4] for multipliers on $H^{p}$ spaces, p＜1.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2000년도 ITC-CSCC -2
• /
• pp.657-660
• /
• 2000

This paper proposes a Data-Flow-Graph (DFG) restructuring to reduce calculating errors in loop folding scheduling. The prime cause of calculating error is rounding errors due to the restriction of the operation digit of functional units. This rounding error is increased more by using multipliers than adders, so reducing the number of multiplications and putting off them as much as possible reduce rounding errors. The proposed approach reduces the number of multiplications by restructuring DFG in loop folding.

• 대한산업공학회지
• /
• 제34권2호
• /
• pp.190-204
• /
• 2008

Recently, it has been strongly required to establish a systematic and sustainable performance investigation and evaluation framework on governmental funding projects for IT small and medium-sized enterprises. In this paper, Data Envelopment Analysis (DEA) models are adopted for performance evaluation on governmental funding projects for IT small and medium-sized enterprises. A new data structure is proposed for the DEA performance evaluation. Generally, in using DEA models, DEA multipliers restriction is critical to achieve the reliability of DEA optimal solutions. Based on the outputs and inputs considered in this study, Acceptance Region (AR) constraints are generated and incorporated into the DEA models so as to improve the reliability of DEA efficiency scores. Associated with AR Type I (AR-I), AR Global Model (ARGM) constraints, DEA/ (AR-I, ARGM) models are designed and then sensitivity analysis follows investigating the robustness of DEA efficiency scores relating to AR constraints adjustment. Finally, a performance evaluation is illustrated regarding governmental direct funding projects from Ministry of Information and Communication (MIC) in Korea where each project unit (i.e. Decision Making Unit (DMU)) is determined whether it is efficient or not. By using DEA/(AR-I, ARGM) models designed in this paper, robustly efficient DMUs are gradually identified according to the successive AR constraints adjustment. Among 25 DMUs, results show that 6 DMUs such as B, E, G, Q, S, Y are determined as robustly efficient against AR constraints intermediate adjustment.