Financial performance optimization is an important embodiment of enterprises to improve operational efficiency and optimize management level. The article proposes a method of financial performance optimization and evaluation using group intelligence algorithm in order to optimize the financial performance of enterprises. EVA is introduced to establish the evaluation index of enterprise financial performance. The financial performance prediction model is constructed according to the propagation process of BP neural network, and the IPSO-BP algorithm is utilized to avoid BP from falling into local optimum and improve the prediction accuracy. In the learning ability test, the relative errors of the EVA value, EVA payoff and EVA rate of the IPSO-BP algorithm are controlled within 6%, 8% and 10% respectively, and the average relative error of the model application results is 3.87%. The model in this paper can achieve more accurate financial performance assessment and prediction, which is conducive to the optimization of financial performance management of enterprises.

The problem of English education quality is worth exploring in depth, and quantifying the indicators of English education can help to understand the problems in teaching and improve the quality of teaching. The study firstly establishes the English education quality evaluation index system, including five first-level indexes of teaching resources, teaching content, teacher quality, teaching effect and teaching quality feedback and 15 second-level indexes, such as network resources, book resources and comprehensive teaching content. On this basis, the combination weights are determined by fusing the G2 method and the projection tracing method through the combination assignment method to eliminate the one-sidedness problem of adopting a single assignment method, and then the cloud model theory is introduced to establish the English education evaluation model based on the cloud model. Problems and shortcomings of multi-objective linear programming weight allocation in English education evaluation system are found through the evaluation results, which lead to low multi objective linear programming weight allocation in English education evaluation system.

In order to enable ships to operate stably for a long time under complex sea conditions, all kinds of ships have an urgent need for gyroscopic rocking reduction devices. This paper takes the double gyro rocking reduction device with better rocking reduction effect as the research object, establishes its corresponding nonlinear dynamic equations, adopts the energy method to establish the differential equations of motion, and deduces the dynamic model of the rocking reduction double gyro. A parameter optimization model is established with the main objective of improving the shaking reduction effect, and the key components of the shaking reduction double gyro are optimized. The bacterial foraging optimization algorithm is selected to solve the model, and the multi-objective parameter optimization model is established. For one to five wave classes, the middle value of the wave height of the meaningful wave is selected for the dynamic simulation experiment of the double gyro. When the wave level is less than three time level, the rocking reduction performance of the rocking reduction double gyro reaches 87.5%, 78.1% and 77.78%, respectively, and the transverse rocking reduction performance is good. Under the simulation environment of sea state I (wave height 2.5m, average period 7s) and sea state II (righteous wave height 2.5m, average period 12s), the rocking reduction efficiencies of the ship after parameter optimization are improved by 6.44% and 10.09%, respectively.

With the rapid development of computer vision technology, image enhancement technology involves an increasingly wide range of research content. At the current stage, picture hierarchy enhancement technology is a research hotspot in the field of image enhancement. This paper proposes an oil painting image enhancement network based on positive probability distribution guidance. The multidimensional spatial information of the samples is obtained through the multibranch information extraction architecture in the network structure, and the probability distribution estimation module estimates the probability distribution through the obtained multidimensional spatial information. In addition, a new image enhancement method based on the RGB color balance method is proposed, which combines the multi-scale Retinex enhancement algorithm with color recovery and the RGB, Lab color space histogram adaptive stretching algorithm, to further improve the effect of oil painting image display. The experimental results show that the method has a better image color bias correction effect compared with the existing techniques. In terms of subjective evaluation, the average subjective score of this paper’s method in three different aesthetic levels reaches 9.15, obtaining a high evaluation. The samples enhanced based on this paper’s algorithm all obtained high aesthetic index scores, indicating that the oil paintings under this paper’s algorithm are in line with the public aesthetics, which is of great significance to the work of oil painting artists.

AI technology can accurately capture and feedback user emotions in digital media interaction to realize precise interaction. In this paper, we design an AI emotion interactivity enhancement model based on multimodal fusion, and apply the neural network model of Bi-GRU and dual attention mechanism to fuse the long and short-term emotion classification results of the tested samples at the decision level to obtain the final emotion classification results. Then the weight coefficient vector of each sentiment category is calculated based on the sentiment classification confusion matrix of the classifier, which is used as the a priori knowledge for multimodal sentiment analysis for decision fusion. The performance is examined on the MOSI dataset and the AI-based interaction design strategy in digital media is proposed. Analyzing the interaction design effect, the interaction design applying the model of this paper has better user experience sense, emotional arousal, pleasure level, and emotional feedback effect in subjectivity evaluation than the control group, and 75% of the experimental subjects think that the feedback-adjusted digital media has a better pleasure level.

In the era of artificial intelligence, human-computer collaborative teaching has become a new picture of future development in the field of education. Based on the theory of human-computer collaboration and the theory of production-oriented approach (POA), this paper constructs a university English POA teaching model based on human-computer collaboration. It also combines the speech recognition algorithm, S-T behavioural analysis method and social network analysis method to conduct a case study on the current situation of college English classroom teaching under this instructional design model. Meanwhile, a teaching experiment is designed to verify the effectiveness of the constructed POA teaching model. The results of the case study show that most of the university English courses favour the lecture mode, with less interaction between students, and the classroom is dominated by teacher lectures and teacher-student interactions, but at the same time, many teachers begin to experiment with the discussion mode, which increases teacher-student interactions and student-student interactions in the classroom. In addition, the experimental group adopts the POA teaching mode and the control group adopts the traditional lecture mode, and its independent samples t-test results show that the experimental group is significantly better than the homogeneous control group in the dimensions of interest, ability, attitude, and test scores in English literacy after the experiment (P<0.05), which suggests that the combination of AI technology and the production-oriented method can effectively improve the effectiveness of the design of university English literacy teaching and achieve better teaching effectiveness and has potential application value.

Market economy is characterized by the uncertainty of supply and demand, so enterprises can realize the optimization of inventory cost control only by reasonably forecasting the demand of supply chain. This paper studies a supply chain demand forecasting method based on machine learning. The factors affecting supply chain demand are collected and analyzed, and the ARMA model, which combines autoregressive model and moving average model, is used to forecast supply chain demand. Then, through the introduction of procurement cost, storage cost and time cost, a multi-level inventory model is established, and the immune genetic algorithm is used to solve the model to find the optimal inventory cost. The experimental results show that the prediction model has good forecasting performance. After using the optimized scheme, the total inventory cost of the enterprise supply chain is reduced by 17.35% and 13.69% respectively. It can be seen that, on the whole, the method in this paper has a good effect of supply chain demand forecasting and cost control.

Service Oriented Architecture (SOA), as a distributed computing architecture, is widely used to build efficient, maintainable and scalable information systems. This paper focuses on SOA design optimization based on reinforcement learning and cloud computing to achieve resource scheduling optimization with a view to improving the service quality of SOA applications. The asynchronous dominant action evaluation algorithm (A3C) based on policy gradient is used as the decision core of the cloud resource scheduler, and the residual recurrent neural network (R2N2) is introduced to construct the cloud resource scheduler based on the A3C-R2N2 algorithm to promote resource scheduling optimization. In the resource scheduling deployment strategy performance experiments, the median average latency of the stochastic dynamic scheduling strategy based on policy gradient learning proposed in this paper is reduced to 9.99% and 56.25% of the direct deployment, respectively, and the CPU utilization rate is also improved by 20.72% compared to the direct deployment. The loss function and reward function of the A3C-R2N2 algorithm in this paper begin to converge after the number of practice reaches 10,000 times and the number of training episodes reaches 300, respectively. Compared with random deployment and nearby deployment strategies, the deployment strategy based on A3C-R2N2 algorithm in this paper has an average service response time of 9.3622s, which is optimal.

Large-span steel structures are prone to wind vibration under wind loads, which affects the safety and performance of the structure, and wind vibration control is the key to its design. This paper takes the large-span steel structure as the research object, firstly introduces the theory and method related to wind vibration control analysis, constructs the topology-optimized inertial capacitance damper controlled wind vibration response dynamic equation of super high-rise building to analyze the influence law of wind speed and wind direction on the dynamic characteristics of the structure, and then further strengthens the vibration control ability of the structure through reasonable arrangement and parameter adjustment. The deformation of ETABS model in y-direction is larger than that in xdirection under 50-year wind load, and the maximum displacements in y- and x-directions are 18.72 mm and 11.65 mm, respectively. The y-direction interstory displacement angle meets the code requirement limit (2.65×10-4). The amplitude of the acceleration time-range curve of its top floor structure is between ±0.08, which meets the requirements for comfort. The optimization of the reinforcement layer using continuum topology optimization is better than the optimization of the optimal location arrangement according to finite element software. The results of node displacements and inter-story displacement angles of each story of the modified structural model under wind load meet the limits of top story displacement and inter-story displacement angle, and the performances are similar to those of the extended-arm truss structural model.

The introduction of performance evaluation in the educational management of colleges and universities is conducive to the formation of result-oriented concepts and management methods of student educational management. In this paper, we select the indicators of educational management conditions, processes and results to design the performance evaluation index system of educational management. Using the hierarchical analysis method, the eigenvectors and maximum eigenvalues are calculated to determine the weights of each index element of the index system. Then apply the gray correlation method to evaluate the educational management performance of the five universities by calculating, one by one, the absolute difference between each indicator sequence (comparative sequence) and the corresponding element of the reference sequence of the object to be evaluated after the data are dimensionless. The analysis found that, according to the formula for calculating the degree of correlation between the actual level of educational management performance and the ideal educational management performance situation, the comprehensive correlation degree of each sample of colleges and universities in the five stages is Z = (0.3333, 0.3951, 0.4600, 0.5031, 0.5946, 1.0000), and the rankings of colleges and universities in terms of the performance of educational management from the highest to the lowest are Academy 4, Academy 2, Academy 5, Academy 3, Academy 1. HEI 3 and HEI 1 should reflect on the shortcomings, enhance the digital construction of teaching informationization, deepen the collaboration between schools and enterprises, and improve the performance of educational management of colleges and universities.