An improved optimization algorithm, namely, multi-strategy-sparrow search algorithm (MSSSA), is proposed to solve highly non-linear optimization problems. In MSSSA, a circle map is utilized to improve the quality of the population. Moreover, the adaptive survival escape strategy (ASES) is proposed to enhance the survival ability of sparrows. In the producer stage, the craziness factor integrated with ASES is introduced to enhance the search accuracy and survival ability. In the scout stage, the ASES facilitates sparrows successful escape from danger. Besides, opposition-based learning or Gaussian–Chachy variation helps optimal individuals escape from local solutions. The performance of the MSSSA is investigated on the well-known 23 basic functions and CEC2014 test suite. Furthermore, the MSSSA is applied to optimize the real-life engineering optimization problems. The results show that the algorithm presents excellent feasibility and practicality compared with other state-of-the-art optimization algorithms.

The prediction technology of remaining useful life has received a lot attention to ensure the reliability and stability of complex mechanical equipment. Due to the large-scale, non-linear, and high-dimensional characteristics of monitoring data, machine learning does not need an exact physical model and prior expert knowledge. It has robust data processing ability, which shows a broad prospect in the field of life prediction of complex mechanical and electrical equipment. Therefore, a remaining useful life prediction algorithm based on Random Forest and Bi-directional Long Short-Term Memory (RF-BiLSTM) is proposed. In the RF-BiLSTM algorithm, RF is utilized to extract health indicators that reflect the life of the equipment. On this basis, a BiLSTM neural network is used to predict the residual life of the device. The effectiveness and advanced performance of RF-BiLSTM are verified in commercial modular aviation propulsion system datasets. The experimental results show that the RMSE of the RF-BiLSTM is 0.3892, which is 47.96%, 84.81%, 38.89%, and 86.53% lower than that of LSTM, SVR, XGBoost, and AdaBoost, respectively. It is verified that RF-BiLSTM can effectively improve the prediction accuracy of the remaining useful life of complex mechanical and electrical equipment, and it has certain application value.

No abstract

Currently, a concern about power resource constraints in the distribution environment is being voiced increasingly, where the increase of power consumption devices overwhelms the terminal load unaffordable and the quality of power consumption cannot be guaranteed. How to acquire the optimal offloading decision of power resources has become a problem that needs to be addressed urgently. To tackle this challenge, a novel reinforcement learning algorithm named Deep Q Network with a partial offloading strategy (DQNP) is proposed to optimize power resource allocation for high computational demands. In the DQNP, a coupled coordination degree model and Lyapunov algorithm are introduced, which trade-offs and decouples the relationships between local-edge and latency–energy consumption. To derive the optimal offloading decision, the resource computation utility function is selected as the objective function. In addition, model pruning is availed to further improve the training time and inference results. Results show that the proposed offloading mechanism can significantly decrease the function value and decline the weighted sum of latency and energy consumption by an average of 3.61%–7.31% relative to other state-of-the-art algorithms. Additionally, the energy loss in the power distribution process is successfully mitigated; furthermore, the effectiveness of the proposed algorithm is also verified.