Optimization Model Of Process Research Articles

Fast and Robust Visual Tracking with Few-Iteration Meta-Learning

Visual object tracking has been a major research topic in the field of computer vision for many years. Object tracking aims to identify and localize objects of interest in subsequent frames, given the bounding box of the first frame. In addition, the object-tracking algorithms are also required to have robustness and real-time performance. These requirements create some unique challenges, which can easily become overfitting if given a very small training dataset of objects during offline training. On the other hand, if there are too many iterations in the model-optimization process during offline training or in the model-update process during online tracking, it will cause the problem of poor real-time performance. We address these problems by introducing a meta-learning method based on fast optimization. Our proposed tracking architecture mainly contains two parts, one is the base learner and the other is the meta learner. The base learner is primarily a target and background classifier, in addition, there is an object bounding box prediction regression network. The primary goal of a meta learner based on the transformer is to learn the representations used by the classifier. The accuracy of our proposed algorithm on OTB2015 and LaSOT is 0.930 and 0.688, respectively. Moreover, it performs well on VOT2018 and GOT-10k datasets. Combined with the comparative experiments on real-time performance, our algorithm is fast and robust.

Process Optimization Models Using Artificial Intelligence and Digital Transformation of The Insurance Industry

Abstract In recent years, there have been fundamental changes worldwide in the field of the business models, in the way these create value and compete, and in the field of services supplied to customers, which have been driven by the way technology is used to interact with them all. All these changes must be supported by the appropriate strategy, a scalable infrastructure and modern technology capable of delivering: target achievement, resilience, scalability, security, information privacy and technological innovation. The insurance industry pioneer’s adaptability to new technologies, considering its clear benefits in terms of improving customer service, mitigating risk and last but not least, increasing the profitability of the business itself. This study aims to analyze how the company which manages the mandatory home insurance system in Romania, Romanian Insurance Pool Against Natural Disaster – PAID S.A has been integrated in this process of digital transformation and to present the circumstances in which the solutions were developed and implemented as well as the effects generated by them.

Molecular reconstruction of vacuum gas oils using a general molecule library through entropy maximization

Vacuum gas oil (VGO) is the most important feedstock for hydrocracking processes in refineries, but its molecular composition cannot be fully acquired by current analysis techniques owing to its complexity. In order to build an accurate and reliable molecular-level kinetic model for reactor design and process optimization, the molecular composition of VGO has to be reconstructed based on limited measurements. In this study, a modified stochastic reconstruction-entropy maximization (SR-REM) algorithm was applied to reconstruct VGOs, with generation of a general molecule library once and for all via the SR method at the first step and adjustment of the molecular abundance of various VGOs via the REM method at the second step. The universality of the molecule library and the effectiveness of the modified SR-REM method were validated by fifteen VGOs (three from the literature) from different geographic regions of the world and with different properties. The simulated properties (density, elemental composition, paraffin-naphthene-aromatics distribution, boiling point distribution, detailed composition of naphthenes and aromatics in terms of ring number as well as composition of S-heterocycles) are in good agreement with the measured counterparts, showing average absolute relative errors of below 10% for each property.

Semi-Physical Simulation Optimization Method for Bolt Tightening Process Based on Reinforcement Learning

Bolted connections are the main method of connecting the components of an aero-engine low pressure rotor. Due to the influence of the elastic interaction relationship, it is easy to cause uneven distribution of the preload force of the bolt group, although it can meet the stiffness needs; however, it may lead to the deflection of the spatial relative position of the components, which is can easily cause coaxiality overrun. In view of the contradictory problem of optimization between coaxiality and stiffness of rotor assembly, this paper proposes a semi-physical simulation optimization method for the bolt tightening process based on reinforcement learning. Firstly, by studying the elastic interaction mechanism between the bolts, the elastic interaction matrix was established using finite element simulation data. On this basis, a coaxiality prediction model for the bolt tightening process was established using a GRU (gate recurrent unit) network to realize the prediction of coaxiality in the bolt tightening process. Then, through the analysis of the bolt connection stiffness, a stiffness calculation model containing the bolt stiffness, the stiffness of the connected parts, and the contact stiffness of the joint surface in series was constructed to realize the calculation of the stiffness during the bolt tightening process. Finally, with the bolt preload force as the optimization variable, coaxiality and stiffness as the optimization target, and the tightening torque and preload force of the installed bolts in the actual assembly process as the constraints, a semi-physical simulation optimization model of the bolt tightening process was established using reinforcement learning to realize the optimization of the bolt tightening process. Moreover, through the semi-physical simulation optimization method, the bolt tightening process can be installed and adjusted at the same time.

Virtual Industrial Process Optimization Models with Cloud Management

The increasing development in ICT has driven and accelerated the advent of IoT technologies, creating the opportunity to empower and innovate data management and production management within factories with data collection, integration, storage and processing as a key challenge. By using the technologies offered by the fourth industrial revolution, it has been possible to test models through simulation before implementing them in real physical systems and with greater complexity. Specifically, by using a combination of cloud computing, IoT platforms, programming software and calculation software, three industrial processes have been virtualized with protocols based on the TCP/IP standard. By considering relevant parameters and variables, an optimization model is developed to improve the operation sequences reducing costs within the factory and a second model to minimize the total cost of operation subject to power balance and production limits. An indispensable element to consider a 4.0 architecture is the database. Mongodb Atlas NoSQL has been able to provide the research with both effective data ingestion and retrieval, which are challenging tasks in case of opting for a SQL database. This paper focuses on the concept of optimization models and their data management from the cloud. The results show that Mongodb is a highly reliable database with enormous scalability and reliability to store data coming from mathematical models or sensors of virtualized production processes, with or without a predefined structure. Moreover, with IoT platforms, it is possible to visualize the retrieved data and interpret them in a graphical way to improve a decision making within the processes.

Inhomogeneous deep Q-network for time sensitive applications

Deep Q-network (DQN) has attracted increasing attention from both industry and academic communities. Existing methods mostly formulate the decision process as discrete agent-environment interactions, while the intervals between successive interactions are largely neglected, which may otherwise reveal important signals in real-world applications. To bridge this gap, this paper proposes to explicitly model the time intervals in DQN. Specifically, we first cast the agent-environment interactions onto a continuous time dimension, and then define a time-aware learning objective and the corresponding Bellman operator. For sample efficient training, we approximate the Q-function with a neural network, where the time information is modeled by the point process. The intensity function in point process and Q-function are seamlessly integrated by sharing the same history summarization module, such that the time interval information can directly influence the model optimization process. To close the gap between the approximated and optimal Q-function, we theoretically analyze the sample complexity of our model by deriving the finite time bound in continuous time. We conduct both simulation and real-world experiments to demonstrate our model's effectiveness.

Current Trends in the Development of Higher Library and Information Education in Germany

The purpose of the article is to explore the experience and identify best practices for modernizing the content and organizational structure of the librarians’ graduate training in Germany.
 The methodology. The research was conducted using a set of theoretical and empirical methods of scientific knowledge. These allowed to identify the main stages of evolution of the higher library education system in Germany, the establishment of classical and innovative models of its modern development. The following scientific approaches were cognitive tools for achieving this goal: system, historical, comparative, as well as research methods: statistical, modelling, content analysis of the 19 German Universities’ 30 educational programs for Bachelors, Masters and Doctors of Philosophy in “Library and Information Sciences”.
 The results. The generalization of the study’s empirical basis allowed us to conclude that 19 German Universities offer more than 30 educational programs to their applicants for “Library and Information Sciences” higher education at various levels, including 45% bachelor’s, 35% master’s, and 20% doctoral degrees. Terms of bachelor’s educational programs are 3–3.5 years, master’s — 2–2.5 years, doctoral — 3–4 years. In the context of the global communication space digitalization the most popular are universal educational programs in “Information Systems”, “Information Science”, and “Information Management”, which train professionals for all types of information institutions and provide extensive opportunities for profiling with a focus on scientific information management or solutions tasks of library, cultural and educational management with a focus on work in school, university and public libraries.
 The topicality. Germany has a long tradition of forming a national book culture and training high-quality librarians, which contributed to the development of one of Europe’s most powerful library, publishing and bookselling industries. Germany’s accession to the Bologna Process made adjustments to the national system of training librarians, identified new vectors of its European integration development, formed modern models for the educational process optimization through the implementation of levelling, succession, flexibility, and interdisciplinarity of education programs in library. In the context of the higher library education system modernization in Ukraine, these innovative practices need to be studied to determine the possibilities of borrowing the most effective of them.
 The practical significance. The results of the analysis of the profiling of modern bachelor’s and master’s degrees in German universities show their interdisciplinarity and focus on training a universal information specialist capable of performing professional tasks in all types of document communication institutions (libraries, archives, media libraries, publishing houses, news agencies, mass media, etc.). Starting with the 3rd semester most bachelor’s degree programs provide their applicants with a choice of a particular specialization, deepening professional competencies through a wide range of optional disciplines and training through participation in the implementation of real interdisciplinary projects. This experience of German colleagues should be borrowed by Ukraine as well.

Two-stage energy scheduling optimization model for complex industrial process and its industrial verification

As complex industries move towards small batches and customization, the uncertainty of the operation time of intermittent electrical devices increases, leading to an increasing fluctuation range of the complex industrial electrical load. To ensure the regular production process, the fossil power plant increases the installed capacity of generators, which can sharply raise carbon emissions. Renewable energy generation has been introduced to reduce carbon emissions and achieve power generation sustainability. However, the instability of renewable energy generation increases the fluctuation range of transmission voltage, leading to highly unsafe electricity transmission. Energy Storage Systems (ESSs) solves the instability problem of renewable energy generation. Thus, this study proposes a two-stage energy scheduling optimization model for complex industrial processes. The first stage proposes a scheduling optimization model for intermittent electrical devices with high electricity consumption. The second stage proposes a scheduling optimization model for the ESS to optimize the transmission capacity proportion of photovoltaic (PV) power plants, ESS, and fossil power plants. The results show that the proposed two-stage scheduling optimization model for the complex industrial process can reduce electricity cost by 7.1%–9.1%, and carbon emissions by 384 tons of standard coal in a year.

Simulation analysis and verification of take-up judder caused by cumulative geometric deviation of clutch friction plates

In this paper, the take-up judder phenomenon caused by the accumulated geometric deviation of clutch plates is simulated and verified by experiments. Firstly, the vibration test and data analysis are carried out for the sample vehicle and related benchmark vehicles, and the vibration target is determined. Then, a simulation analysis and optimization model of the clutch engagement process in the vehicle take-up process was established, and the optimal control target was determined by multi-condition analysis of the accumulation geometric deviation factors of the clutch. Finally, several clutch deviation samples are verified, and total deviation of clutch control strategy is given. This paper provides a quantitative analysis method of the clutch cumulative geometric deviation in the take-up process, which provides a reference for the clutch parameter setting and the vibration control of the vehicle in the take-up process.

Optimasi Suhu dan Waktu Proses Delignifikasi pada Isolasi Selulosa dari Tongkol Jagung

This study aims to optimize the temperature and time of the delignification process to isolate cellulose from corn cobs. The research was carried out in stages, namely preparing corn cob powder samples and the delignification process. Statistical regression modeling and optimization of the temperature and time of the delignification process used the response surface (RSM) methodology. Central Composite Design (CCD) was used for experimental design and analysis of the effect of temperature and time of the delignification process on lignin, cellulose, hemicellulose, and fiber yields. The optimum process conditions were determined using the desirability method. The analysis of the diversity of the quadratic polynomial model shows that the temperature and time of the deligninification process significantly affect changes in the characteristics of lignin, cellulose, hemicellulose, and fiber yields. The optimal condition of the corncob fiber delignification process was obtained at a temperature of 100.38? for 79.17 minutes. It produced the characteristics of delignified fiber with lignin, cellulose, hemicellulose, and yields of 6.47%, 73.73%, 5, respectively. 39% and 57.76% w/w. The optimization model is valid with a desirability value of 0.71. Thus, the temperature and time optimization model of the delignification process can be applied in real conditions to produce delignified corncob fiber with optimum characteristics.

A hybrid multi-objective bi-level interactive fuzzy programming method for solving ECM-DWTA problem

Electronic countermeasure (ECM) has become one of the most significant factors in modern warfare, in the course of combat, the electronic jamming allocation tasks need to be flexibly adjusted with the change of combat stage, which puts forward higher requirements for the modeling and solution method of this kind of problems. To solve the ECM dynamic weapon target assignment (ECM-DWTA) problem, a hybrid multi-target bi-level programming model is established. The upper level takes the sum of the electronic jamming effects in the whole combat stage as an optimization objective, and locally optimizes the ECM weapon (ECM-WP) assignment scheme in each stage. The lower level takes the importance expectation value of the target subjected to interference and combat consumption as double optimization objectives to globally optimize the ECM-WP assignment scheme. Focus on solving this complex model, a hybrid multi-objective bi-level interactive fuzzy programming algorithm (HMOBIF) is proposed, in this method, exponential membership function is used to describe the satisfaction degree of each level. When solving the multi-objective optimization problem composed of membership functions in the upper and lower levels, we use the MOEA/D algorithm to obtain the Pareto Front (PF) solution set, and then each solution in PF is evaluated and selected by the TOPSIS multi-criteria evaluation method. This local and global interactive optimization process of bi-level model is actually the process of executing observation-orientation-decision-action loop in practical combat. According to the current example, we conduct numerical simulation on the parameters in the model and obtain the parameter values suitable for the model solution. The computational experiments on different scale ECM-DWTA problems show that HMOBIF method is superior to four bi-level programming algorithms in terms of performance index, and can better solve ECM-DWTA problems.

Research on Optimization of Food Industry Processing Process Based on Computational Intelligence

Computational intelligence refers to an algorithm inspired by the body structure of natural animals and plants and unique landscapes. It can perform simultaneous multithreaded work. In the processing process of the food industry, multiple factors must be considered at the same time to optimize various parameters simultaneously. Therefore, this research combines computational intelligence algorithms with the processing flow of the food industry, conducts research on the optimization of food processing systems and processes, and fully understands the needs of the market, the competitiveness of the company’s products, and the people facing marketing. The MA-CI model was used to optimize the entire process of food processing and production, and multiple variables were integrated. The following conclusions were drawn: (1) When the node size = 100 , calculate the optimal value of the intelligent model MA − CI = 0.3196 , the worst value = 0.3120 , and the average plus − minus variance = 0.3155 ± 4.6278 × 10 − 6 , in each index. Both are better than the Greedy algorithm model and the EA model. When the node size = 200 , the optimal value of the calculation intelligence model MA − CI = 0.3050 , the worst value = 0.2900 , and the average plus and minus variance = 0.2970 ± 8.9251 × 10 − 6 ; when the node size = 300 , the calculation intelligence model (MA-CI) optimal value = 0.2608 , the worst value = 0.2499 , and the mean plus − minus variance = 0.2551 ± 1.0028 × 10 − 5 ; when the node size = 500 , calculate the optimal value of intelligent model MA − CI = 0.2857 , the worst value = 0.2812 , and the mean plus and minus variance = 0.2834 ± 2.8230 × 10 − 6 . (2) In the electronic circuit network, calculate the intelligent model (MA-CI): Rc = 0.3174 , NMI = 0.7388 , Δ Ec = 0 , and k = 14.43 ; in the USAir network, the computational intelligence model (MA-CI) is as follows: Rc = 0.3041 , NMI = 0.8632 , Δ Ec = 0 , and k = 7.43 ; the computational intelligence model (the uniqueness and optimality of MA-CI) has resulted in the optimal solution for the optimization of the global machining process. (3) Using six online databases to predict and verify the MA-CI model for process optimization in the food industry, it was found that the MA-CI model for process optimization in the food industry was better than the standard value. (4) Perform performance testing on the optimized MA-CI model. It is found that under the MA-CI model, MUTAG = 98.2 ± 4.3 , ENZYMES = 96.2 ± 4.3 , PTC = 85.2 ± 4.3 , PROTEINS = 82.0 ± 3.2 , NCI 1 = 80.2 ± 2.0 , and D & D = 91.9 ± 0.5 , which are better than the other models. The optimized MA-CI model can control the profit problem in food processing and production, and it is found that the MA-CI model can make the enterprise obtain the maximum profit in the shortest period.

Experimental validation of an adsorbent-agnostic artificial neural network (ANN) framework for the design and optimization of cyclic adsorption processes

The efficacy of an adsorbent agnostic machine-learning surrogate model for rapid design and optimization of a Skarstrom cycle vacuum swing adsorption (VSA) process is experimentally validated. The surrogate model is trained to predict the process performance using adsorbent features that include hypothetical Langmuir adsorption isotherm parameters, particle density, porosity and bed voidage, and process variables such as pressure, step duration and feed velocity. The training data was generated from a detailed process model for 20,000 unique combinations of the training variables. The model shows high accuracy of Radj2>0.99 for predicting key performance parameters such as product purity, recovery and productivity. The ability of this surrogate to predict the experimental performance for the purification of O2 from air on two adsorbents, namely 13X and LiX zeolites, was studied. Two separate multi-objective optimization studies, to maximize purity and recovery, and to maximize productivity and purity were performed. For these optimization studies, the volumetrically measured isotherms of N2 and O2 were used as inputs to the surrogate model. Note that these isotherms were not a part of the dataset used to train the model. Nine points were chosen from the Parteo curves and the corresponding decision variables were used as set-points in a two-column lab-scale rig. The average difference between the calculated and experimentally measured purity, recovery and productivity was 3%, 5% and 9%, respectively. This study provides the necessary confidence to use surrogate-based process models for adsorbent screening and adsorption process optimization.

Artificial neural network based prediction of ultimate buckling strength of liquid natural gas cargo containment system under sloshing loads considering onboard boundary conditions

In this study, the ultimate buckling strength of a GTT NO96 liquid natural gas (LNG) cargo containment system under a sloshing impact load was investigated while considering the design environmental conditions thereof, such as the cryogenic temperature and flexible boundary conditions owing to the inner hull structures. As the dynamic buckling capacities of LNG cargo containment boxes show a large nonlinearity against several design parameters, we applied an artificial neural network to generalize the numerical analysis results and develop design ultimate capacities of the system for early design purposes. Through one-hot encoding, a character variable was converted into a variable form suitable for the artificial neural network, and a prediction model optimized for the current system was constructed through a model optimization process. A comparison of the design values with finite element (FE) simulation results shows a very good agreement. Finally, we provide a prediction model: a GitHub repository, BUNO96, which is suitable for the calculation of ultimate buckling strength of NO96 box.

Research on Optimization of Process Parameters of Traditional Chinese Medicine Based on Data Mining Technology.

Data mining technology and methods are used to effectively optimize manufacturing process parameters due to the complexity and uniqueness of the process parameters. The data-mining-based optimization method for traditional Chinese medicine (TCM) process parameters is presented, along with a list of process parameters that have shown to be effective in actual production. The influencing factors of process parameters are analyzed and modeled using an attribute weight analysis and classification analysis algorithm. The optimization scheme of process parameters that meet the requirements is selected, and an example is given for verification, by selecting data records that fall within a certain error range and incorporating the rules of association knowledge discovery. The support vector classification algorithm has a higher accuracy, despite the algorithm's results being understandable. The support vector regression algorithm developed a reliable process optimization model.

Process optimization of high machining efficiency and low surface defects for HSD milling UD-CF/PEEK with limited thermal effect

Carbon fiber reinforced polyetheretherketone (CF/PEEK) is widely applied in aerospace and medical fields due to its excellent strength-to-weight ratio, fatigue resistance and biocompatibility. However, traditional dry machining significantly limits the improvement of machining efficiency and surface quality. Addressing this issue, this paper proposes a process optimization method of high machining efficiency and low surface defects for high-speed dry (HSD) milling UD-CF/PEEK, based on analyzing the thermal effect of cutting temperature. Mixed-level orthogonal experiments are designed for dry milling unidirectional (UD) CF/PEEK laminates with the fiber orientation of 0° and 90°. Three-dimensional surface roughness Sq, cutting temperature Tc and material removal rate MRR are extracted to characterize experiment results. The results show that the cutting temperature is mainly affected by cutting speed and fiber orientation, and machined surface defect is mainly caused by the thermal effect of high cutting temperature. Then, a process optimization model of high machining efficiency and low surface defects is established based on the genetic algorithm optimized BP (GA-BP) neural network, non-dominated sorting genetic algorithm (NSGA-II) and technique for order preference by similarity to an ideal solution (TOPSIS). The optimization objectives are Sq, MRR and Tc, and Tc also serves as a constraint of the thermal effect. Verification experiments are carried out, and results reveal that the surface defects are significantly reduced by limited thermal effect, and the model can be used to effectively improve machining efficiency and surface quality for HSD milling UD-CF/PEEK. Besides, the material removal rate generated by the HSD milling operation can easily reach 2–3 times higher than the traditional low-speed dry milling operation. The optimized cutting speed in HSD milling UD-CF/PEEK is recommended as 1300–1600 m/min. This study provides significant technical guidance for the machinability improvement of HSD milling UD-CF/PEEK in practice.

Supercritical CO2 extraction of lipids and astaxanthin from Atlantic shrimp by-products with static co-solvents: Process optimization and mathematical modeling studies

Shrimp processing by-products are a source of valuable biomaterials/bioactive materials (such as lipids and astaxanthin). Extraction of these compounds could not only decrease environmental impacts associated with shrimp harvesting but also result in economic benefits. This work investigates the caonditions for supercritical extraction combined with co-solvents for the extraction of lipids and astaxanthin from Atlantic shrimp by-products (Pandalus borealis). Extraction conditions (pressure, carbon dioxide flow, and temperature) impact lipid and astaxanthin yields and compositions. A model was developed to describe the SC-CO2 extraction rates of lipids/astaxanthin using the Goto et al. model.In experiments using only CO2, the highest lipid yield was extracted at 50 °C and 30 MPa and the maximum astaxanthin yield and total carotenoid content at 60 °C and 32 MPa. Lipid/astaxanthin recovery increased with an increase in pressure; however, the impact of temperature was more complex. Overall extraction rates of lipid/astaxanthin were controlled by the strong solid–solute interaction. Pure SC-CO2 extract had high percentages of neutral lipids but low phospholipids. In co-solvent experiments (using ethanol, 40:60 vol% hexane/acetone, fish oil, or sunflower oil), polar organic co-solvents showed higher lipid/astaxanthin yield and phospholipids/saturated fatty acid concentrations in the extract relative to pure SC-CO2 extract. There was a corresponding decrease in monounsaturated and polyunsaturated fatty acid concentrations in the extract. Sunflower oil as a co-solvent showed higher astaxanthin, sterol, and free fatty acid concentrations compared to the fish oil extract.

Cellulose Nanocrystals from Office Paper Waste for Green Mortar: Process Optimization Modeling, Characterization, and Mechanical Properties

Cellulose Nanocrystals from Office Paper Waste for Green Mortar: Process Optimization Modeling, Characterization, and Mechanical Properties

Improving Productivity of Misaligned Boring of Hydraulic Cylinder Holes with a Rotating Cutter Block Based on Computer Simulation Results

The analysis of existing processing methods was carried out. A series of virtual experiments was performed; based on their results dependences of the height of residual ridges, the temperature on the workpiece surface and the temperature in the cutting zone on the parameters of the processing mode when boring internal holes with a rotating cutter block were constructed based on the results of virtual computer simulation. An optimization model of the boring process was built. The objective function was received. Within the optimization model framework, solutions were obtained that allow, at the highest value of axial feed, to ensure the required surface quality and avoid changes in the physical and mechanical properties of the material.

Geospatial model of COVID-19 spreading and vaccination with event Gillespie algorithm

We have developed a mathematical model and stochastic numerical simulation for the transmission of COVID-19 and other similar infectious diseases that accounts for the geographic distribution of population density, detailed down to the level of location of individuals, and age-structured contact rates. Our analytical framework includes a surrogate model optimization process to rapidly fit the parameters of the model to the observed epidemic curves for cases, hospitalizations, and deaths. This toolkit (the model, the simulation code, and the optimizer) is a useful tool for policy makers and epidemic response teams, who can use it to forecast epidemic development scenarios in local settings (at the scale of cities to large countries) and design optimal response strategies. The simulation code also enables spatial visualization, where detailed views of epidemic scenarios are displayed directly on maps of population density. The model and simulation also include the vaccination process, which can be tailored to different levels of efficiency and efficacy of different vaccines. We used the developed framework to generate predictions for the spread of COVID-19 in the canton of Geneva, Switzerland, and validated them by comparing the calculated number of cases and recoveries with data from local seroprevalence studies.