Parallel Path Strategy Theory and Numerical Illustrations

Abstract

This paper deals with parallel-planning strategies for the development of a new item. Let us assume that several proposals for developping the new item are available. Usually one is interested to choose the “best” of these proposals with respect to some criterion. But it may happen that the data available at this point of the development process are very inaccurate and uncertain and therefore unreliable. For example a proposal promising low costs and low time to bring the project to the end may turn out to be very expensive after its completion. In such situations a wrong decision may be avoided by pursuing a parallel path approach. Several, say m, of the proposals are pursued to a review-point at which point the best project is selected and brought to completion while the other still remaining approaches are stopped. This is the simplest model of a parallel path-approach; generalization to several review points, however, are possible and done in the literature (see Marschak [3]).

This work was done, when the author was a research worker at “Studiengruppe für Systemforschung”, D 69 Heidelberg 1, Werderstr. 35