Polygonal Approximation and Reduction of Activity Resource Consumption Distributions in Program Management Networks.
Abstract
The Air Force, as well as the other agencies of the Department of Defense, has had limited success with the use of networks in both analysis and program management applications. A practical methodology for approximation network resource consumption, other than simulation, would increase the flexibility of network utilizations. This study investigates the feasibility of using polygonal (straight-line) approximation to characterize the distribution of activity resource consumption in program management networks. Curvilinear functions (normal and exponential probability distributions) and uniform probability distributions are approximated by piecewise polygonal functions through the use of Simple Linear Regression. The distribution approximations are then manipulated to approximate the distributions of convolutions, or additive combinations, and maximum operations required in network reductions. The resultant distribution describes or characterizes, the total estimated resources consumption of the network. Results for the probability distributions investigated indicate the Polygonal Approximation and Reduction Technique (PART) is comparable in accuracy to the Q-GERT simulation model.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 01, 1978
- Accession Number
- ADA059311
Entities
People
- Jerry A. Fergeson
- Patallen H. Shortell
Organizations
- Air Force Institute of Technology