PROGRAMMING UNDER UNCERTAINTY: AN EXPERIMENTAL CODE FOR THE COMPLETE PROBLEM,
Abstract
It is very convenient from a computational point of view to assume that the coefficients of a linear programming problem are exactly known. However, the resulting solution can differ widely from the true optimum because some or all of these coefficients may, in fact, be random variables. The 'complete' problem in programming under uncertainty takes into account the possibility of overor under-estimating the requirements and penalizes the over- or under-supply. This paper describes the problem, presents two examples, outlines the mathematical aspects of an algorithm (given in AD-608 990) and finally gives the write-up and the features of an experimental code written for the IBM 7090 or 7094. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1964
- Accession Number
- AD0611844
Entities
People
- David Kohler
- Roger J-B Wets
Organizations
- Boeing