TARGET AGGREGATION OF SECOND-ORDER MULTI-AGENT SYSTEMS WITH SWITCHING INTERCONNECTION

In this paper, the target aggregation is investigated for a multi-agent system consisting of second-order agents and multiple leaders. Sufficient conditions are proposed to make the agents approach the target set spanned by these moving leaders. With the condition on switching interconnection topologies, all mobile agents can asymptotically track the dynamical target set specified by multiple moving leaders with bounded error. Moreover, discussion on the case with static leaders is also given.

Game-theoretical explorations of the mesoscale flow structure and regime transitions in bubble columns

Understanding the mesoscale structure and regime transition in bubble columns is of great significance for reactor design and scaleup.Based on the energy-minimization multiscale(EMMS)model,a noncooperative game model with constraints is proposed to investigate the structural properties of gas-liquid systems in which small and large bubbles are chosen as players and the energy consumption form the objective function.The conservation equations of the system can be regarded as the constraints of the game.For the formulated noncooperative game model,the concept of the generalized Nash equilibrium(GNE)is used to characterize the solution.An algorithm is developed to numerically compute the GNE and some important structural parameters in the system.The numerical results show the existence of the GNE for all values of the superficial gas velocity Ug.As Ug varies,the trends in the state variables can be observed and the critical point of Ug identified.The overall trend of the flow regime transition agrees with the original EMMS model and experimental results,although the GNE calculation also reveals different single-bubble dominant mechanisms with increasing Ug.

The Game Theoretical Approach for Multi-phase Complex Systems in Chemical Engineering

This paper explores the application of noncooperative game theory together with the concept of Nash equilibrium to the investigation of some basic problems on multi-scale structure, especially the meso-scale structure in the multi-phase complex systems in chemical engineering. The basis of this work is the energy-minimization-multi-scale （EMMS） model proposed by Li and Kwauk （1994） and Li, et al. （2013） which identifies the multi-scale structure as a result of ＇compromise-in-competition between dominant mechanisms＇ and tries to solve a multi-objective optimization problem. However, the existing methods often integrate it into a problem of single objective optimization, which does not clearly reflect the ＇compromise-in-competition＇ mechanism and causes heavy computation burden as well as uncertainty in choosing suitable weighting factors. This paper will formulate the compromise in competition mechanism in EMMS model as a noncooperative game with constraints, and will describe the desired stable system state as a generalized Nash equilibrium. Then the authors will investigate the game theoretical approach for two typical systems in chemical engineering, the gas-solid fluidiza- tion （GSF） system and turbulent flow in pipe. Two different cases for generalized Nash equilibrinm in such systems will be well defined and distinguished. The generalize Nash equilibrium will be solved accurately for the GSF system and a feasible method will be given for turbulent flow in pipe. These results coincide with the existing computational results and show the feasibility of this approach, which overcomes the disadvantages of the existing methods and provides deep insight into the mechanisms of multi-scale structure in the multi-phase complex systems in chemical engineering.

Class conditional distribution alignment for domain adaptation

In this paper,we study the problem of domain adaptation,which is a crucial ingredient in transfer learning with two domains,that is,the source domain with labeled data and the target domain with none or few labels.Domain adaptation aims to extract knowledge from the source domain to improve the performance of the learning task in the target domain.A popular approach to handle this problem is via adversarial training,which is explained by the H△H-distance theory.However,traditional adversarial network architectures just align the marginal feature distribution in the feature space.The alignment of class condition distribution is not guaranteed.Therefore,we proposed a novel method based on pseudo labels and the cluster assumption to avoid the incorrect class alignment in the feature space.The experiments demonstrate that our framework improves the accuracy on typical transfer learning tasks.

Distributed stochastic mirror descent algorithm for resource allocation problem

In this paper,we consider a distributed resource allocation problem of minimizing a global convex function formed by a sum of local convex functions with coupling constraints.Based on neighbor communication and stochastic gradient,a distributed stochastic mirror descent algorithm is designed for the distributed resource allocation problem.Sublinear convergence to an optimal solution of the proposed algorithm is given when the second moments of the gradient noises are summable.A numerical example is also given to illustrate the effectiveness of the proposed algorithm.

Distributed best response dynamics for Nash equilibrium seeking in potential games

In this paper,we consider distributed Nash equilibrium(NE)seeking in potential games over a multi-agent network,where each agent can not observe the actions of all its rivals.Based on the best response dynamics,we design a distributed NE seeking algorithm by incorporating the non-smooth finite-time average tracking dynamics,where each agent only needs to know its own action and exchange information with its neighbours through a communication graph.We give a sufficient condition for the Lipschitz continuity of the best response mapping for potential games,and then prove the convergence of the proposed algorithm based on the Lyapunov theory.Numerical simulations are given to verify the resultandillustrate the effectiveness of the algorithm.

Variational Bayesian data analysis on manifold

In this paper, variational inference is studied on manifolds with certain metrics. To solve the problem, the analysis is first proposed for the variational Bayesian on Lie group, and then extended to the manifold that is approximated by Lie groups. Then the convergence of the proposed algorithm with respect to the manifold metric is proved in two iterative processes： variational Bayesian expectation （VB-F） step and variational Bayesian maximum （VB-M） step. Moreover, the effective of different metrics for Bayesian analysis is discussed.

Distributed solver for linear matrix inequalities: an optimization perspective

In this paper,we develop a distributed solver for a group of strict(non-strict)linear matrix inequalities over a multi-agent network,where each agent only knows one inequality,and all agents co-operate to reach a consensus solution in the intersection of all the feasible regions.The formulation is transformed into a distributed optimization problem by introducing slack variables and consensus constraints.Then,by the primal–dual methods,a distributed algorithm is proposed with the help of projection operators and derivative feedback.Finally,the convergence of the algorithm is analyzed,followed by illustrative simulations.

Distributed Communication-Sliding Mirror-Descent Algorithm for Nonsmooth Resource Allocation Problem

This paper considers a distributed nonsmooth resource allocation problem of minimizing a global convex function formed by a sum of local nonsmooth convex functions with coupled constraints.A distributed communication-efficient mirror-descent algorithm,which can reduce communication rounds between agents over the network,is designed for the distributed resource allocation problem.By employing communication-sliding methods,agents can find aε-solution in O(1/ε)communication rounds while maintaining O(1/ε^(2))subgradient evaluations for nonsmooth convex functions.A numerical example is also given to illustrate the effectiveness of the proposed algorithm.