Search

In this paper, a distributed nonsmooth nonconvex optimization (DNNO) problem with affine inequality and nonsmooth convex inequality constraints is studied. A continuous-time distributed neurodynamic algorithm is proposed to solve this problem. Under the assumed conditions, for any initial state, the solution of distributed neurodynamic algorithm is bounded and globally exists, and will converge to the critical point set of distributed problems in a finite time. Compared with other DNNO algorithms, distributed neurodynamic algorithm has a lower dimension and does not need to satisfy the assumption that the feasible region is bounded. Finally, a series of numerical examples are given to verify the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

• A distributed algorithm is proposed to solve graph semi-supervised learning problem by leveraging the graph topology. • The convergence of the distributed algorithm is explicitly proved. • Numerical results verify the good performance and low cost of the distributed algorithm. Graph semi-supervised learning (GSSL) plays an important role in data classification by leveraging the similarity across the graph topology and convex optimization with Laplacian-based regularization. However, the current algorithm to solve the problem is centralized approach calling for heavy computational cost, particularly when the data is of large volume. In this paper, an innovate distributed algorithm is proposed to solve the problem, which is based on the decomposition of the similar graph. Contrary to the centralized approach, the distributed algorithm only requires the neighboring information for solving the optimization. It is proved that difference between the solutions of the distributed algorithm and centralized counterpart is upper bounded. We apply the proposed algorithm to both the synthetic and real-world datasets. The numerical results verify the effectiveness of the proposed distributed algorithm. [ABSTRACT FROM AUTHOR]

Multi-agent reinforcement learning (MARL) has been successfully applied in many fields. In MARL, the policy evaluation problem is one of crucial problems. In order to solve this problem, distributed Temporal-Difference (TD) learning algorithm is one of the most popular methods in a cooperative manner. Despite its empirical success, however, the theory of the adaptive variant of distributed TD learning still remain limited. To fill this gap, we propose an adaptive distributed temporal-difference algorithm (referred to as MS - ADTD) under Markovian sampling over time-varying networks. Furthermore, we rigorously analyze the convergence of MS - ADTD , the theoretical results show that the local estimation can converge linearly to the optimal neighborhood. Meanwhile, the theoretical results are verified by simulation experiments. • This paper proposes a distributed adaptive TD learning algorithm under Markovian sampling. • It theoretically analyzes the non-asymptotic convergence performance of the proposed algorithm. • It verifies the performance of the proposed algorithm by experiments. [ABSTRACT FROM AUTHOR]

• The data-driven inverse optimization corrects the DR execution online. • The proposed strategy is realized in a distributed manner and free of initialization. • This paper provides rigorous stability proof for the proposed optimization. • D-PPDS algorithm solves the DR optimization with global inequality constrain. This paper provides an insight into the demand response (DR) optimization in distribution markets consisting of a retailer and multiple demand response aggregators (DRA), where a retailer determines DR incentives based on power consumption profile. Conventional DR optimization with global states and constraints is intractable to be implemented in a distributed framework, which restricts the application feasibility and the potential profit of DR. To handle these limitations, we design a multi-agent architecture for distributed demand response (DDR). An online data-mining method is developed to identify the characteristics of DR. A leader–follower structure decomposes the original problem into a leader problem with global variables and aggregators of sub-problems, where discrete singular consensus is designed to broadcast the leader's strategy to followers in real-time. The distributed perturbation primal–dual sub-gradient (D-PPDS) algorithm is proposed to solve the DDR problem with global inequality constraints in a completely distributed fashion. The proposed DDR strategy is tested by an actual case. The simulation results demonstrate that the asynchronous D-PPDS algorithm can obtain the near-optimal solution of the problem with global inequality constraints, and is robust against delay or plug-and-play. [ABSTRACT FROM AUTHOR]

• Calculation speed, information privacy, and multiple objectives are considered. • Layering, multi-objective economic dispatch and distributed algorithm are combined. • A framework of multi-layer distributed multi-objective economic dispatch is proposed. • The multi-layer distributed multi-objective consensus algorithm is proposed. • The proposed algorithm has higher information privacy and faster calculation speed. Traditional multi-objective economic dispatch problems all apply centralized multi-objective optimization methods. However, with the rapid development of smart grids, issues such as system calculation speed, robustness, and information privacy have been increasingly vitally. Although the distributed multi-objective optimization method can solve the problems of system robustness and privacy when the scale of the interconnected power system expands and the number of agents increases, each agent independently optimizes its sub-problems and then exchanges information to complete the entire system optimization. The pressure of system communication calculation leads to reduce the overall calculation speed greatly. In response to the above problems, this paper proposes a multi-layer distributed multi-objective consensus algorithm. This method first calculates the optimal power generation of each area of each layer through the network topology and then calculates the power of each unit in each area in parallel according to the calculated optimal power generation. The analysis of the three system simulation results of IEEE118-bus, IEEE2154-bus, and Xinjiang Turpan 117-bus shows that the proposed method can solve the problems of multi-objective and information privacy, and realize the rapid solution of economic dispatch in large-scale multi-area interconnected power systems. [ABSTRACT FROM AUTHOR]

This paper presents a novel control and planning algorithm for cooperative slung-load transportation using multiple quadrotors. Each aerial robot modeled as a quadrotor is connected to a non-uniform payload by the cables. Udwadia-Kalaba equations are used for calculating the forces on the cables acting on each robot. A novel leaderless distributed adaptive control algorithm is presented which not only ensures the safe transport of payload in the generated path in the presence of external disturbances but also manages to solve the problems related to the interaction of payload with the aerial robots. Moreover, the payload dynamics is not considered in the design of the controller which facilitates the load transportation in any size, shape, or weight by the robotic team without compromising performance. Then, an improved rapidly exploring random tree star (RRT*) algorithm is developed to generate desired path for each aerial robot in order to avoid known obstacles for cooperative payload transportation problem. The proposed path planning algorithm is capable of finding a free collision path in any situation, specially where high maneuverability is crucial to avoid collision of the payload with obstacles. Three numerical simulations are conducted to verify the proposed planning and control scheme. Additionally, in the distributed control algorithm simulation, two indexes including ISE and IADU demonstrate the improved performance of the distributed control over independent control of robots. The superior performance of the proposed controller in terms of reduced tracking errors in presence of external disturbances is also shown in comparison with a distributed nonlinear H ∞ design. [ABSTRACT FROM AUTHOR]