A probabilistic model for planning research and development under linear constraints

For a complicated R&D system described by a determinate network of arbitrary topology with more than one initial vertex and a single terminal vertex, the paper considers the problem of its management by the criterion of maximal subjective probability of technical success of a set of developments which is defined by the worst value among the technological chains forming a project under given budgetary and temporal constrains and linear relations between the variables. The described formulation is of great applied importance for managing R&D projects that are efficiently described in network terms.