Abstract The slime mould algorithm (SMA), as an emerging and promising swarm intelligence algorithm, has been studied in various fields. However, SMA suffers from issues such as easily getting trapped in local optima and slow convergence, which pose challenges when applied to practical problems. Therefore, this study proposes an improved SMA, named HESMA, by incorporating the covariance matrix adaptation evolution strategy (CMA-ES) and storing the best position of each individual (SBP). On one hand, CMA-ES enhances the algorithm’s local exploration capability, addressing the issue of the algorithm being unable to explore the vicinity of the optimal solution. On the other hand, SBP enhances the convergence speed of the algorithm and prevents it from diverging to other inferior solutions. Finally, to validate the effectiveness of our proposed algorithm, this study conducted experiments on 30 IEEE CEC 2017 benchmark functions and compared HESMA with 12 conventional metaheuristic algorithms. The results demonstrated that HESMA indeed achieved improvements over SMA. Furthermore, to highlight the performance of HESMA further, this study compared it with 13 advanced algorithms, and the results showed that HESMA outperformed these advanced algorithms significantly. Next, this study applied HESMA to five engineering optimization problems, and the experimental results revealed that HESMA exhibited significant advantages in solving real-world engineering optimization problems. These findings further support the effectiveness and practicality of our algorithm in addressing complex engineering design challenges.

Abstract Robotic manipulation in surgical applications often demands the surgical instrument to pivot around a fixed point, known as remote center of motion (RCM). The RCM constraint ensures that the pivot point of the surgical tool remains stationary at the incision port, preventing tissue damage and bleeding. Precisely and efficiently controlling tool positioning and orientation under this constraint poses a complex Inverse Kinematics (IK) problem that must be solved in real-time to ensure patient safety and minimize complications. To address this problem, we propose PivotIK, an efficient IK solver that combines efficient evolutionary exploration with multi-objective Jacobian-based optimization. PivotIK can track desired tool poses accurately while satisfying the RCM constraint in real-time. We evaluated PivotIK through simulations and real-world experiments using redundant robotic manipulators with multi-degree-of-freedom surgical instruments. We compare PivotIK with other IK solvers in terms of solve rates, computation times, tracking errors, and RCM errors under various scenarios, including unconstrained and RCM-constrained trajectories. Our results show that PivotIK achieves superior performance, solving the IK problem in less than 1 ms with errors below 0.01 and 0.1 mm for tracking and RCM, respectively. Our real-world experiments confirm the effectiveness of PivotIK in ensuring smooth trajectory tracking and RCM compliance. PivotIK offers a promising solution for real-time IK for robotic manipulation under RCM constraints in surgical applications.

Abstract Most traditional methods for solving partial differential equations (PDEs) require the costly solving of large linear systems. Neural operators (NOs) offer remarkable speed-ups over classical numerical PDE solvers. Here, we conduct the first exploration and comparison of NOs for three-dimensional topology optimization. Specifically, we propose replacing the PDE solver within the popular Solid Isotropic Material with Penalization (SIMP) algorithm, which is its main computational bottleneck. For this, the NO not only needs to solve the PDE with sufficient accuracy but also has the additional challenge of providing accurate gradients which are necessary for SIMP’s density updates. To realize this, we do three things: (i) We introduce a novel loss term to promote gradient-consistency. (ii) We guarantee equivariance in our NOs to increase the physical correctness of predictions. (iii) We introduce a novel NO architecture called U-Net Fourier neural operator (U-Net FNO), which combines the multi-resolution properties of U-Nets with the Fourier neural operator (FNO)’s focus on local features in frequency space. In our experiments we demonstrate that the inclusion of the novel gradient loss term is necessary to obtain good results. Furthermore, enforcing group equivariance greatly improves the quality of predictions, especially on small training datasets. Finally, we show that in our experiments the U-Net FNO outperforms both a standard U-Net, as well as other FNO methods.

Abstract Most research on collective robotic construction (CRC) separates the architectural design and robotic path planning phases of the overall construction process. Specifically, a structure is designed and afterwards sent to a planner or compiler that returns instructions for the assembly of the structure with the robots at hand. Although this has led to the assembly of spatially complex structures, it obscures the planning process, making it inaccessible to the architect. Considering that one potential of CRC is that the architect can perform as more than a designer of static structures, this paper showcases how agent-based modelling can collapse the architectural design and robotic planning phases for CRC. As such the overall construction workflow is upended, leading to more designer control, adjustment for tolerances in the construction process, a more general understanding of the processes, and the potential for architectural reconfiguration when working with CRC systems. This is demonstrated through the presentation of an agent-based model for assembling a planar structure using a previously developed CRC system.

Abstract With the widespread use of computer-aided technologies like CAD/CAM/CAPP in the product manufacturing process, a large amount of process data is constantly generated and data-driven process planning has shown promising potentials for effectively reusing the process knowledge. However, a lot of labeled data are needed to train a deep learning model for effectively extracting the embedded knowledge and experiences within these process data and the labeling of process data is quite expensive and time-consuming. This paper proposes a cost-effective process design intents extraction approach for process data by combining active learning (AL) and self-paced learning (SPL). First, the process design intents inference model based on Bi-LSTM is generated by using a few pre-labeled samples. Then, the prediction uncertainty of each unlabeled sample is calculated by using a Bayesian neural network, which can assist in the identification of high confidence samples in SPL and low confidence samples in AL. Finally, the low confidence samples with manual-labels and the high confidence samples with pseudo-labels are incorporated into the training data for retraining the process design intents inference model iteratively until the model attains optimal performance. The experiments demonstrate that our approach can substantially decrease the number of labeled samples required for model training and the design intents in the process data could be inferred effectively with dynamically undated training data.

Abstract The current study on the synthesis problems of four-bar mechanism trajectories primarily relies on establishing a numerical atlas based on trajectory characteristics and employing neural networks to synthesize mechanism parameters. However, this approach has several shortcomings, including a vast database, inefficient retrieval, and challenges in maintaining accuracy. This paper presents a method for synthesizing a trajectory-generation mechanism that combines the extreme gradient boosting (XGBoost) algorithm with a genetic algorithm (GA). The purpose is to synthesize, based on a particular trajectory, the dimensions and installation position parameters of a four-bar mechanism. The paper classifies the trajectories according to their shape features and geometric center placements, thereby improving the accuracy of the XGBoost model for synthesizing mechanisms. The XGBoost algorithm is employed to synthesize the basic dimensional parameters for the mechanism, with the relative slopes of trajectories as input features. The synthesized basic dimensional parameters are turned into parameters for the actual mechanism by researching the scaling, translation, and rotation relationships between mechanisms and the trajectories they generate. The accuracy of the generated trajectories from the synthesized mechanism can be improved by applying GA to optimize the mechanism parameters. Five comparative examples are provided in this research for the different scenarios of given trajectory curves and trajectory points. The effectiveness and accuracy of the proposed approach in this study are validated in comparison to existing research methods by comparing errors between the generated trajectories and the given trajectories.

Abstract The study of variety of Newtonian nanofluids subject to various physical model parameters gained much interest of engineers and scientists. Owing to their coolant and absorption characteristics, these are broadly found in chemical engineering, biomedical engineering (expansion and contraction of veins and arteries), detection of cancer cells through magnetic nanoparticles, microchips, and particularly in petroleum industry. This study focuses on investigation of nanofluid heat transfer applications inside a channel formed by expanding/contracting walls. A new heat transport model is introduced by adding the effects of nanoparticles and molecular diameters, thermal radiations, and walls permeability. Then, numerical code for the model is developed and executed to analyze the dynamics of the model from physical aspects. For expanding (${\alpha }_1 = 1.0,2.0,3.0,4.0$) and contracting (${\alpha }_1 = - 1.0, - 2.0, - 3.0, - 4.0$) walls, the velocity is examined maximum in the channel center. However, the fluid movement in the working domain is in reverse proportion for ${Re} = 1.0,3.0,5.0,7.0$. Further, high absorbent walls (${A}_1 = 0.1,0.3,0.5,0.7$) controlled the motion for both ${\alpha }_1 > 0$ and ${\alpha }_1 < 0$, respectively. The addition of thermal radiation number ${Rd} = 0.1,0.3,0.5,0.7$ played the role of catalytic parameter which imperatively increased the fluid temperature. Further, contracting walls and temperature ratio number ${\theta }_r = 0.1,0.3,0.5,0.7$ reduced the temperature and this decrease is rapid in conventional fluid.

Abstract In the realm of medical diagnosis, recent strides in deep neural network-guided magnetic resonance imaging (MRI) restoration have shown promise. Nevertheless, persistent drawbacks overshadow these advancements. Challenges persist in balancing acquisition speed and image quality, while existing methods primarily focus on singular tasks like MRI reconstruction or super-resolution (SR), neglecting the interplay between these tasks. To tackle these challenges, this paper introduces the mutual co-attention network (MCAN) specifically designed to concurrently address both MRI reconstruction and SR tasks. Comprising multiple mutual cooperation attention blocks (MCABs) in succession, MCAN is tailored to maintain consistency between local physiological details and global anatomical structures. The intricately crafted MCAB includes a feature extraction block, a local attention block and a global attention block. Additionally, to ensure data fidelity without compromising acquired data, we propose the channel-wise data consistency block. Thorough experimentation on the IXI and fastMRI dataset showcases MCAN’s superiority over existing state-of-the-art methods. Both quantitative metrics and visual quality assessments validate the enhanced performance of MCAN in MRI restoration. The findings underscore MCAN’s potential in significantly advancing therapeutic applications. By mitigating the trade-off between acquisition speed and image quality while simultaneously addressing both MRI reconstruction and SR tasks, MCAN emerges as a promising solution in the domain of magnetic resonance image restoration.

Abstract In the steel stockyard of the shipyard, the sorting work to relocate the steel plates already stacked to retrieve the target steel plate on the fabrication schedule is labor-consuming work requiring the operation of overhead cranes. To reduce the sorting work, there is a need for a method of stacking the steel plates in order of fabrication schedules when the steel plates arrive at the shipyard from the steel-making companies. However, the conventional optimization algorithm and heuristics have limitations in determining the optimal stacking location of steel plates because the real-world stacking problems in shipyards have vast solution space in addition to the uncertainty in the arrival order of steel plates. In this study, reinforcement learning is applied to the development of a real-time stacking algorithm for steel plates considering the fabrication schedule. Markov decision process suitable for the stacking problem is defined, and the optimal stacking policy is learned using an asynchronous advantage actor-critic algorithm. The learned policy is tested on several problems by varying the number of steel plates. The test results indicate that the proposed method is effective for minimizing the use of cranes compared with other metaheuristics and heuristics for stacking problems.

Abstract We present a bee foraging behavior-driven mutational salp swarm algorithm (BMSSA) based on an improved bee foraging strategy and an unscented mutation strategy. The improved bee foraging strategy is leveraged in the follower location update phase to break the fixed range search of salp swarm algorithm, while the unscented mutation strategy on the optimal solution is employed to enhance the quality of the optimal solution. Extensive experimental results on public CEC 2014 benchmark functions validate that the proposed BMSSA performs better than nine well-known metaheuristic methods and seven state-of-the-art algorithms. The binary BMSSA (bBMSSA) algorithm is further proposed for feature selection by using BMSSA as the selection strategy and support vector machine as the classifier. Experimental comparisons on 12 UCI datasets demonstrate the superiority of bBMSSA. Finally, we collected a dataset on the return-intentions of overseas Chinese after coronavirus disease (COVID-19) through an anonymous online questionnaire and performed a case study by setting up a bBMSSA-based feature selection optimization model. The outcomes manifest that the bBMSSA-based feature selection model exhibits a conspicuous prowess, attaining an accuracy exceeding 93%. The case study shows that the development prospects, the family and job in the place of residence, seeking opportunities in China, and the possible time to return to China are the critical factors influencing the willingness to return to China after COVID-19.