Search

The scarcity of in-situ ocean observations poses a challenge for real-time information acquisition in the ocean. Among the crucial hydroacoustic environmental parameters, ocean sound velocity exhibits significant spatial and temporal variability and it is highly relevant to oceanic research. In this study, we propose a new data-driven approach, leveraging deep learning techniques, for the prediction of sound velocity fields (SVFs). Our novel spatiotemporal prediction model, ST-LSTM-SA, combines Spatiotemporal Long Short-Term Memory (ST-LSTM) with a self-attention mechanism to enable accurate and real-time prediction of SVFs. To circumvent the limited amount of observational data, we employ transfer learning by first training the model using reanalysis datasets, followed by fine-tuning it using in-situ analysis data to obtain the final prediction model. By utilizing the historical 12-month SVFs as input, our model predicts the SVFs for the subsequent three months. We compare the performance of five models: Artificial Neural Networks (ANN), Long Short-Term Memory (LSTM), Convolutional LSTM (ConvLSTM), ST-LSTM, and our proposed ST-LSTM-SA model in a test experiment spanning 2019 to 2022. Our results demonstrate that the ST-LSTM-SA model significantly improves the prediction accuracy and stability of sound velocity in both temporal and spatial dimensions. The ST-LSTM-SA model not only accurately predicts the ocean sound velocity field (SVF), but also provides valuable insights for spatiotemporal prediction of other oceanic environmental variables. [ABSTRACT FROM AUTHOR]

U-shaped assembly lines have been popularly adopted in electronics and appliances to improve their flexibility and efficiency. However, most past studies assumed that the processing time of each task is fixed and hence just considered the task allocation but ignored worker assignment. In this paper, the processing time of each task depends on the workers and then the cooperative optimization of task allocation and workers assignment is considered in U-shaped assembly line balancing problems to optimize the cycle time. Later, an enhanced migrating birds optimization algorithm (EMBO) is proposed to solve it. In the EMBO algorithm, since this new problem has two subproblems: task allocation and worker assignment, the prevent work designs two neighborhood structures to improve the leader and following birds. Furthermore, the temperature acceptance criteria, to judge whether the neighbor replaces current following bird, are developed to ensure the diversity of population and avoid being trapped in the local optimum. And a competitive mechanism is introduced to increase the probability of the promising birds locating in the front of the line. The proposed algorithm is compared with other well-known algorithms in the literature, and the numerical results demonstrate that the proposed algorithm outperforms other algorithms. [ABSTRACT FROM AUTHOR]

The collaborative robots (cobots) are increasingly being utilized in industries due to the advancement in the field of robotic technology and also due to the increase in labor costs. The cobots on the assembly line can be utilized to complete the tasks independently or assist the workers to complete the tasks. This study considers the U-shaped assembly line balancing problem with cobots, where several cobots with different purchasing costs are selected under the budget constraint. Three mixed-integer programming models are formulated to optimize the cycle time, and the built models are capable of solving the small-sized instances optimally. Two algorithms, artificial bee colony algorithm and migrating bird optimization algorithm, are developed and improved to tackle the large-sized instances, where new encoding scheme and decoding procedure are developed for this new problem. The computational tests demonstrate that the utilization of collaborative robots reduces the cycle time effectively in the assembly line. The comparative study on a set of instances shows that the proposed methodologies obtain competing performance in comparison with other 12 implemented algorithms. [ABSTRACT FROM AUTHOR]

Industries are increasingly looking for opportunities at utilizing collaborative robots in assembly lines to perform the tasks independently or assist the human workers due to the advancement of industry 4.0 technologies. Purchasing cost is one of the important factors to be considered by production managers, while designing or redesigning assembly line when collaborative robots are being utilized. Several objectives are to be optimized in an assembly line balancing problem and optimizing line efficiency along with purchasing cost sometimes results in conflicting situation. This paper presents the first study to tackle the cost-oriented assembly line balancing problem with collaborative robots, where several different types of collaborative robots with different purchasing costs are available and selected. A multi-objective mixed-integer programming model is developed to minimize the cycle time and the total collaborative robot purchasing cost. The multi-objective migrating bird optimization algorithm is developed to obtain a set of high-quality Pareto solutions. This algorithm utilizes the fast non-dominated sorting approach to update the population and develops a restart mechanism to select one solution in the permanent Pareto archive to replace the abandoned solution which remains unchanged for several iterations. The computational study validates that the utilization of the multi-objective model is reasonable and developed algorithm produces competing performance in comparison with multi-objective non-dominated sorting genetic algorithm II, multi-objective simulated annealing algorithm and two multi-objective artificial bee colony algorithms. [ABSTRACT FROM AUTHOR]

Seafloor real-time positioning is important for the instantaneous detection of seafloor crustal motion, seismic activities, hydrological rapid variations and rapid geodetic datum updates. Current GNSS-Acoustic (GNSS-A) seafloor positioning usually utilizes batch processing of long-term observations in the postprocessing mode. Seafloor real-time positioning can be achieved using sequential processing of the epoch-wise observations. We propose the sequential GNSS-A seafloor point positioning method with modeling of the sound speed variation in the kinematic survey. We propose real-time modeling of the sound speed variation using oceanography analysis data and we then calculate the random walk (RW) process noise of the acoustic nadir total delay (NTD) caused by the sound speed variation. The experiments conducted in the South China Sea and Japan Trench validate the method performance in terms of epoch-wise positioning accuracy, high-resolution sound speed variation, and filter convergence time. The difference between the estimated sound speed and the in situ sound velocity profiles was 0.128 m/s root mean square. The vessel track of the line and circle combination performs best with a high positioning accuracy and a short convergence time. The position in real-time sequential processing with the modeled NTD RW process noise converged to a 3D range of 0.125 m from the static post-determined position. The a posteriori residual of the acoustic travel time observations was equivalently 0.270 m in range. These findings can improve the temporal resolution of the GNSS-A positioning and oceanography. [ABSTRACT FROM AUTHOR]

Exact methods have shown advanced and promising performance in solving the simple assembly line balancing problem, known as NP-hard. This research investigates the impact of various structural parameters on the performance of exact methods, including branching methods, search direction, method to achieve upper bounds, utilized lower bounds, utilized dominance rules and search strategy. In accordance with the structural parameter evaluation, utilized dominance rules and search strategy have shown the most important effect on the exact methods' performance. This research also improves and re-implements three well-known exact methods [i.e., SALOME, bounded dynamic programming (BDP) heuristic and branch, bound and remember (BBR) algorithm] using effective parameters. Computational study demonstrates that the utilization of high-performance structural parameters enhances the performance of exact methods by a significant margin. The re-implemented BBR method with proper parameters shows clear superiority over all the published exact methods and might be regarded as the state-of-the-art exact methodology. [ABSTRACT FROM AUTHOR]

This paper investigated the quality of site-specific surface temperature and surface pressure data in the coastal regions of China, which were interpolated from the European Centre for Medium-Range Weather Forecast Interim reanalysis (ERA-Interim), Japanese 55-year Reanalysis Project (JRA-55), and the National Centers for Environmental Prediction Final (NCEP FNL) reanalysis surface meteorological datasets as well as from the new Global Pressure and Temperature (GPT2) model data. The measured temperature and pressure along with the collocated GPS data from 2014 were collected from 25 observation stations evenly located in the region. Compared with the actual meteorological observations, the performances of the interpolated data from three reanalysis datasets differ marginally, with the root mean square errors (RMSEs) of the interpolated surface temperature and pressure less than 2.4 K and 1.6 hPa, respectively; however, the RMSEs of the surface temperature and pressure interpolated from the GPT2 model were 3.0 K and 4.2 hPa, respectively. Data based on GPS PWV products that used the meteorological parameters interpolated from three reanalysis data were very close to those of meteorological observations, with biases within ± 0.4 mm and RMSEs below 0.5 mm in most areas, and the RMSE of PWV using the GPT2 model interpolation data was superior by 2 mm. The measurement of GPS PWV using the interpolated reanalysis meteorological data also compared well with radiosonde observations, with RMSE between them tending to increase with a decrease of the GPS station's latitude. However, the GPS PWV based on the interpolated data could not reflect the true change in water vapor during typhoon events. [ABSTRACT FROM AUTHOR]

Robotics are extensively utilized in modern industry to replace human labor and achieve high automation and flexibility. In order to produce large-size products, two-sided assembly lines are widely applied, where robotics can be employed to operate tasks on workstations. Since the applied traditional optimization methods are limited, the current work presented a new discrete cuckoo search algorithm to solve the two-sided robotic assembly line balancing problem. The original cuckoo search algorithm was modified by employing neighbor operations. Furthermore, a new procedure to generate individuals to replace the abandoned nests was developed to enhance the intensification. Since the considered problem has two subproblems, namely the robot allocation and assembly line balancing, the present work extended the cuckoo search algorithm to cooperative coevolutionary paradigm by dividing the cuckoos into two sub-swarms, each addressing a subproblem. In order to emphasize the exploration, a restart mechanism was employed. The proposed discrete algorithm’s evolution process and convergence were compared with another two popular optimization algorithms, namely the genetic algorithm and particle swarm optimization algorithm. Computational study on the proposed algorithms and other five recent algorithms along with statistical analysis demonstrated that the proposed methods yielded promising results. [ABSTRACT FROM AUTHOR]

An improved ant colony optimization (ACO), namely, station ant colony optimization (SACO), is proposed to solve the type 2 assembly line balancing problem (ALBP-2). In the algorithm, ACO is employed to search different better combinations of tasks (component solutions) for each station; an iteration compress mechanism is proposed to reduce the searching space of feasible solutions of ALBP-2. Three heuristic factors [i.e., (1) task time, (2) number of successors, and (3) number of releasable successors], two pheromones, and a task assignment mechanism are proposed to search better component solutions for every station. Finally, the effectiveness and stability of SACO are confirmed through comparison with literatures in 23 instances included in nine examples. [ABSTRACT FROM AUTHOR]