未知干扰下多智能体系统任意预设时间的滑模编队优化OA

For leader-less multi-agent systems,the problem of distributed formation optimization in predefined time under unknown disturbances was studied,and the global cost function composed of local strongly convex functions for all agents was minimized.A class of formation optimization algorithms based on the sliding mode control was proposed to realize the formation control of multi-agent systems within the predefined time.The al-gorithm was divided into 3 parts:firstly,the integrated sliding mode control strategy was used to guide each a-gent to approach the sliding mode surface in the predefined time,and the external interference was effectively suppressed;then,the design protocol control was employed to guide each agent state to the minimum point of its local cost function;finally,the leaderless formation was realized for all agents to reach the minimum point of the global cost function.The algorithm does not require agents to share the gradients and Hessian matrix in-formation of neighbors,thus saving the information exchange cost,and can deal with highly nonlinear multi-valued strongly convex cost functions.Several examples of numerical experiments demonstrate the effectiveness and reliability of the design control protocol algorithm.