SIMPLE STOCHASTIC NETWORKS: SOME PROBLEMS AND PROCEDURES.
CARNEGIE-MELLON UNIV PITTSBURGH PA MANAGEMENT SCIENCES RESEARCH GROUP
Pagination or Media Count:
The paper provides an analysis of stochastic networks, that is networks whose underlying link times are expressed as random variables. Analytic and Monte Carlo methods for finding the distribution function or parameters thereof of the maximal time through such a network are discussed. The first section, which considers analytical procedures, emphasizes networks whose link times are given by probability functions from an exponential family. It is shown that the Markov property of the exponential distribution may be utilized to simplify analytic computations. In the second section Monte Carlo techniques are described which provide useful information on project completion time distributions. It is argued that these easily applied techniques are more economical fewer computations are required than naive simulation. Throughout the paper, we suggest a blending of the analytical and simulation procedures into an efficient overall method for studying stochastic networks. Author
- Administration and Management
- Operations Research