An algorithm which protects consumers’ privacy is considered for a tatonnement model which determines the equilibrium price such that demand and supply coincide. In this model, consumers bid their demands and a firm bids her supply for the price given by an auctioneer, which is iterated until the total demands and the supply are balanced. Instead of bidding the demands exactly, noise is added to these values, which introduces differential privacy into the iterative algorithm. A definition of adjacency between two privacy information is provided and sensitivity of the algorithm is investigated. The main advantage of the algorithm is to protect consumers’ privacy and to guarantee that the price converges to the equilibrium one in a probabilistic sense.
This paper discusses gain-scheduled state feedback synthesis problems for Markovian jump systems with some time-varying parameters that are available online. In particular, we consider those systems with such time-varying parameters not only in the coefficient matrices but also in the mode transition probability matrices. This paper discusses the stability analysis and synthesis methods of the gain-scheduled state feedback under such a situation that these time-varying parameters are exploited. The effectiveness of the methods is demonstrated through numerical examples.
This paper discusses the robustness of the performance of the closed loop system for a homogeneous finite time proportional-derivative control (FT-PD control) of a one-link robot manipulator. We show that FT-PD control's disturbance rejection performance of low frequency is superior to the conventional PD control via the describing functions estimated by an experiment.