Fully Probabilistic Design for Stochastic Discrete System with Multiplicative Noise

In this paper, a novel algorithm based on fully probabilistic design (FPD) is proposed for a class of linear stochastic dynamic processes with multiplicative noise. Compared with the traditional FPD, the new procedure is presented to deal with multiplicative noise and the system parameters are estimated online using linear optimisation methods. The performance index is characterised by the Kullback-Leibler divergence (KLD) distance. The generalised probabilistic control law is obtained by solving a generalised Riccatti equation that takes the multiplicative noise into consideration. To demonstrate the effectiveness of the proposed method, a numerical example is given and the results are compared with the traditional FPD.