3 results about "Criss-cross algorithm" patented technology

The invention discloses a crisscross optimization robust model predictive control method for an automatic sweeping robot. The crisscross optimization robust model predictive control method is adaptive to a trajectory tracking system with parameter uncertainty. According to the method, a discrete time controlled system with parameter uncertainty, a reference trajectory system and a tracking error system are constructed, performance evaluation indexes are defined, the stable control gain range of a basic RMPC controller is determined by solving a minimum and maximum optimization problem and linear matrix inequality constraints, and a multi-objective optimization function is combined in the range to achieve the stable control of the RMPC controller. The optimal additional control gain is iteratively searched by means of a crisscross algorithm, an optimized controller is formed, the stability and the dynamic performance are considered, the influence caused by parameter uncertainty can be overcome, the rising time and the adjusting time of trajectory tracking are remarkably shortened, the tracking error and the energy consumption are reduced, and the tracking efficiency is improved. And the dual requirements of the automatic sweeping robot on real-time response and control precision are met.

The invention discloses a power grid equipment communication state evaluation method and device, electronic equipment and a storage medium, and the method comprises the steps: constructing an initial BP neural network, obtaining a training sample, setting R stages, and setting T iterations for each stage; using the training samples as particles in the population, and initializing the particles in the population; when single iteration is carried out in each stage, the fitness of each particle in the initial population is solved; screening and updating a particle optimal solution and a global optimal solution through a crisscross algorithm; when the traversal of the R stages is completed, taking a final global optimal solution as a network parameter of the BP neural network; and performing model training on the BP neural network based on the training sample to obtain an evaluation model, and evaluating the communication state of the power grid equipment by adopting the evaluation model. The convergence speed of the algorithm is accelerated, the global search capability is improved, the network parameters of the BP neural network can be quickly obtained, the evaluation work efficiency is high, and the accuracy is high.

The invention discloses a hybrid renewable energy system power dispatching method based on dynamic stability constraint and an adaptive algorithm. The method comprises the following steps: establishing a security constraint unit combination objective function; inputting a conventional security constraint unit commitment constraint condition into the security constraint unit commitment target function, and constructing a hybrid renewable energy system and unit commitment dynamic coupling evaluation model; based on a stability margin function of a critical mode, setting a dynamic stability constraint condition for the evaluation model, performing partition processing, constructing a multi-objective function of a unit commitment under the dynamic stability constraint, and calculating a stability margin of an initial unit commitment scheduling scheme; and adopting an adaptive power variable step size crisscross algorithm, adjusting unit output based on the participation activeness, updating a scheduling scheme and a stability margin of an initial unit combination, calculating a corresponding power generation cost, and performing power scheduling according to the power generation cost. According to the invention, the stability margin can be obviously improved.