32 results about "Evolutionary learning" patented technology

An evolutionary learning method for intelligent monitoring of tool state, using three-way acceleration sensors and microphones to collect vibration signals and acoustic signals, smoothing the signals, and dividing them into training sets and test sets; using stacked autoencoders to analyze dynamic signals Automatically extract deep-level features and classify the extracted features; assign weights to each algorithm according to the accuracy of the training set model, obtain the final predicted tool state through weighted average, and save the relevant parameters of the model; the actual processing process The real-time vibration signal and acoustic signal are input into the saved monitoring model after data preprocessing, and the tool state of the corresponding signal is obtained, the data label with a high level of confidence is saved, and the network parameters are updated, so as to realize the intelligent monitoring of the tool state evolutionary learning. This method can avoid manual participation, reduce computational complexity, and can weaken the impact of machine tool performance degradation on the prediction accuracy of the tool state monitoring model.

The invention discloses a channel network steady-state evolution game method in a blockchain environment. Based on bounded rationality and the "replication dynamic" mechanism of the biological evolution process, the node transactions in the channel network are modeled as a dynamic evolution game model. The model design A dynamic defense mechanism of channel network based on evolutionary game. This mechanism adds a defense strategy to the strategy space of type I nodes, and considers the cost and success rate of type II node attacks under the condition of deploying defense strategies. This defense mechanism can help Type I nodes resist attacks from Type II nodes with a certain probability. The nodes of this model have the evolutionary learning ability of bounded rationality, and can dynamically adjust their respective strategies according to the attacker's different attack strategies to achieve effective defense.

The invention discloses a multi-scale geological feature detection and fusion method based on deep learning and evolutionary learning, and relates to the technical field of geological feature detection and fusion, which includes the following steps: S1, output of reservoir evaluation parameters, S1.01, according to the input Multiple data volumes, seismic data, well logging data generate single and multi-body joint feature data, and mark the feature data. The multi-scale geological feature detection fusion method based on deep learning and evolutionary learning uses its own learning methods and data structures for different data, and realizes different evolution and learning according to different data. Different models can be used for different geology. During the learning process, the self-correction and improvement of the model and the evaluation of parameters will be realized. The learning process does not contain fixed models and models, and the final generated reservoir parameters have more geological significance. For the reservoir parameters Predictions will also be closer to reality.

The invention discloses a digital image filter circuit design method based on FPGA evolutionary learning, which utilizes the optimization characteristics of gene expression for circuit coding and (2+)ES evolutionary strategy, and uses multiple The target model evolves a set of appropriate relational operations, and the designed filter circuit can achieve the best possible image filtering effect. In the learning stage, the filter can obtain an optimized logical composition structure after a limited evolution algebra. Through VHDL conversion and Competition and risk elimination design, realize the hardware circuit of image filtering on FPGA chip. The non-linear filter circuit evolved by the present invention makes the filtered image clear and the edges clear.

The invention discloses a UAV-assisted edge computing method considering user mobility, including: establishing a UAV-assisted computing network driven by a digital twin; and using the UAV-assisted computing network to feed back users, UAVs and Real-time data, location and status of idle devices; with the goal of reducing the total user delay, use distributed deep learning to solve the user's unloading strategy; use differential evolution learning to solve the location deployment of drones. The present invention can effectively use the twin network to feed back user real-time data, obtain a global near-optimal solution UAV deployment and unloading strategy, and reduce user processing delay.

The invention discloses a distributed resource dynamic allocation method based on an evolutionary game theory, which comprises the following steps: S1, judging whether supply and demand of a physicalmachine and a user task are matched, if yes, executing the step S2, and if not, executing the step S3; S2, performing a game between the physical machine and the user task based on the price bargaining strategy of the fraudulent sound potential behavior, judging whether the transaction between the physical machine and the user task can be successful, and if yes, executing the step S4; otherwise, executing the step S3; S3, acquiring a new physical machine or user task, and executing the step S1; and S4, determining a resource exchange price ratio, and allocating resources to the user task. According to the method, a bargaining phenomenon commonly existing in real life is added in a game link; according to the method, the fitness function is combined, so that the main body repeatedly plays agame with the purposes of resource balance and efficient resource utilization in the game link, the strategy of individuals with higher fitness in similar main bodies is learned in the evolutionary learning link, and the excellent performance of the group in the aspects of configuration efficiency, resource balance and the like is continuously promoted.