Stochastic Modeling of Dynamic Value Networks

The proposed project deals with modeling and simulation of dynamic value networks, which are increasingly replacing the traditional vertically integrated enterprises. The project´s objective is to develop a stochastic model of the dynamic behavior of a value network, which would account for both the formation of a functioning network out of available nodes and the changes in performance of individual nodes over time. The proposed approach to modeling combines Ulf Grenander´s general pattern theory with stochastic extension of Jay Forrester´s system dynamics. Estimation of the network state and parameter values is formulated as the Bayesian inference problem, whose numerical solution is assumed using a sequential Monte Carlo algorithm. The model is generic enough to cover value networks in diverse industries. The project explores various definitions of the ´value´concept and the relationship between the value transformation over a specific node and the node performance characteristics.