Sorting Operations Research Articles

Object Detection Post Processing Accelerator Based on Co-Design of Hardware and Software

Deep learning significantly advances object detection. Post processes, a critical component of this process, select valid bounding boxes to represent the true targets during inference and assign boxes and labels to these objects during training to optimize the loss function. However, post processes constitute a substantial portion of the total processing time for a single image. This inefficiency primarily arises from the extensive Intersection over Union (IoU) calculations required between numerous redundant bounding boxes in post processing algorithms. To reduce these redundant IoU calculations, we introduce a classification prioritization strategy during both training and inference post processes. Additionally, post processes involve sorting operations that contribute to their inefficiency. To minimize unnecessary comparisons in Top-K sorting, we have improved the bitonic sorter by developing a hybrid bitonic algorithm. These improvements have effectively accelerated the post processing. Given the similarities between the training and inference post processes, we unify four typical post processing algorithms and design a hardware accelerator based on this framework. Our accelerator achieves at least 7.55 times the speed in inference post processing compared to that of recent accelerators. When compared to the RTX 2080 Ti system, our proposed accelerator offers at least 21.93 times the speed for the training post process and 19.89 times for the inference post process, thereby significantly enhancing the efficiency of loss function minimization.

Sorting Numbers in Ascending Order using Java

Sorting is a fundamental operation in programming that is widely used across various applications, from data analysis to organizing user inputs. This article explores the implementation of sorting numbers in ascending order using java programming language as Java's Arrays.sort().Through code examples and performance analysis, readers will gain an understanding of how different programming languages handle sorting operations, and which languages offer the best performance for specific scenarios. This comparative study aims to provide developers with the knowledge to choose the most appropriate sorting technique for their projects.

A Tomato Recognition and Rapid Sorting System Based on Improved YOLOv10

In order to address the issue of time-consuming, labor-intensive traditional industrial tomato sorting, this paper proposes a high-precision tomato recognition strategy and fast automatic grasping system. Firstly, the Swin Transformer module is integrated into YOLOv10 to reduce the resolution of each layer by half and double the number of channels, improving recognition accuracy. Then, the Simple Attention Module (SimAM) and the Efficient Multi-Scale Attention (EMA) attention mechanisms are added to achieve complete integration of features, and the Bi-level Routing Attention (BiFormer) is introduced for dynamic sparse attention and resource allocation. Finally, a lightweight detection head is added to YOLOv10 to improve the accuracy of tiny target detection. To complement the recognition system, a single-vertex and multi-crease (SVMC) origami soft gripper is employed for rapid adaptive grasping of identified objects through bistable deformation. This innovative system enables quick and accurate tomato grasping post-identification, showcasing significant potential for application in fruit and vegetable sorting operations.

An engineered trafficking biosensor reveals a role for DNAJC13 in DOR downregulation.

Trafficking of G protein-coupled receptors (GPCRs) through the endosomal-lysosomal pathway is critical to homeostatic regulation of GPCRs following activation with agonist. Identifying the genes involved in GPCR trafficking is challenging due to the complexity of sorting operations and the large number of cellular proteins involved in the process. Here, we developed a high-sensitivity biosensor for GPCR expression and agonist-induced trafficking to the lysosome by leveraging the ability of the engineered peroxidase APEX2 to activate the fluorogenic substrate Amplex UltraRed (AUR). We used the GPCR-APEX2/AUR assay to perform a genome-wide CRISPR interference screen focused on identifying genes regulating expression and trafficking of the δ-opioid receptor (DOR). We identified 492 genes consisting of both known and new regulators of DOR function. We demonstrate that one new regulator, DNAJC13, controls trafficking of multiple GPCRs, including DOR, through the endosomal-lysosomal pathway by regulating the composition of the endosomal proteome and endosomal homeostasis.

An Improved Negotiation-Based Approach for Collecting and Sorting Operations in Waste Management and Recycling

This paper addresses the problem of optimal planning for collection, sorting, and recycling operations. The problem arises in industrial waste management, where distinct actors manage the collection and the sorting operations. In a weekly or monthly plan horizon, they usually interact to find a suitable schedule for servicing customers but with a not well-defined scheme. We proposal an improved negotiation-based approach using an auction mechanism for optimizing these operations. Two interdependent models are presented: one for waste collection by a logistics operator and the other for sorting operations at a recycling plant. These models are formulated as mixed-integer linear programs where costs associated with sorting and collection are to be minimized, respectively. We describe the negotiation-based approach involving an auction where the logistics operator bids for collection time slots, and the recycling plant selects the optimal bid based on the integration of sorting and collection costs. This approach aims to achieve an optimization of the entire waste management process. Computational experiments are presented.

Machine Learning-based Differentiation of Japanese Plum Cultivars Based on Metabolite Profiling

This study investigates the metabolite diversity of twelve Japanese plum cultivars grown in South Africa to understand their differential organoleptic characteristics and nutritional properties. The goal is to differentiate or associate these plum cultivars based on their metabolic profiles. Metabolite profiling was conducted using gas chromatography-mass spectrometry (GC-MS) at different postharvest ripening stages. Different unsupervised machine learning algorithms were applied: hierarchical clustering, K-means clustering, Density-Based Spatial Applications with Noise, and principal component analysis (PCA). Results revealed that each cultivar contains a unique combination of 13 amino acids, 4 sugars (contributing to organoleptic characteristics), and numerous phenolic compounds and antioxidant activities (contributing to nutritional value). The levels of these compounds are cultivar-dependent and vary with postharvest stages. The number of clusters of plum cultivars varied with both the clustering algorithm and postharvest stages. However, certain cultivars were consistently grouped regardless of the clustering method, indicating similar characteristics and responses to storage and shelf-life conditions. Similar consistent groupings were observed after cold storage and shelf life. Our findings also showed that K-means clustering is the most effective model for plum cultivar differentiation based on the Silhouette Score and Davies-Bouldin Index. This study enhances our understanding of how metabolites evolve over different postharvest stages and provides a robust framework for differentiating plum cultivars, which can aid in sorting and grading operations. The research offers actionable insights to improve postharvest handling and storage practices, which are critical for maintaining the nutritional quality of plums, an important fruit for human health.

Using Machine Learning to Identify Diseases and Perform Sorting in Apple Fruit

Fruit diseases play a major role in global agriculture, leading to substantial crop losses and influencing food production and economic stability. In this age of Industry 4.0 the fruit sorting is an important part in the food processing wherein this work plays a vital role. In this study, a solution for the detection and classification of apple fruit diseases is proposed and experimentally validated. Deep learning models offer promise for automating disease identification using fruit images, but encounter obstacles such as therequirement for extensive training data, computational complexity, and the risk of overfitting. This study introduces an innovative convolutional neural network (CNN) architecture aimed at addressing these challenges by incorporating a reduced number of layers, thus alleviating computational burdens while maintaining performance. Additionally, augmentation techniques such as shift, shear, scaling, zoom, and flipping are employed to diversify the training set without additional image acquisition. Our CNN model is specifically trained to identify common apple crop diseases like Scab, Rot, and Blotch. Rigorous experimental evaluation demonstrates the effectiveness ofour model, achieving a remarkable classification accuracy of 95.37%. Significantly, our model demonstrates reduced storage requirements and faster execution times compared to existing deep CNN architectures, enabling deployment on handheld devices and resource-limited environments. While other CNN models may offer similar accuracy levels, our approach emphasizes efficiency and resource optimization, rendering it practical for real-world applications in agriculture. Furthermore, our CNN model exhibits resilience to environmental variations and imaging parameters, enhancing its applicability across diverse agricultural settings. By leveraging advanced machine learning techniques, the approach developed in this experimental work contributes to modernizing fruits and vegetables sorting operations in food processing, crop management practices thus promoting agricultural sustainability. The scalability and portability of our model make it suitable for deployment in both small-scale farms and large-scale agricultural operations.

StopThePop: Sorted Gaussian Splatting for View-Consistent Real-time Rendering

Gaussian Splatting has emerged as a prominent model for constructing 3D representations from images across diverse domains. However, the efficiency of the 3D Gaussian Splatting rendering pipeline relies on several simplifications. Notably, reducing Gaussian to 2D splats with a single viewspace depth introduces popping and blending artifacts during view rotation. Addressing this issue requires accurate per-pixel depth computation, yet a full per-pixel sort proves excessively costly compared to a global sort operation. In this paper, we present a novel hierarchical rasterization approach that systematically resorts and culls splats with minimal processing overhead. Our software rasterizer effectively eliminates popping artifacts and view inconsistencies, as demonstrated through both quantitative and qualitative measurements. Simultaneously, our method mitigates the potential for cheating view-dependent effects with popping, ensuring a more authentic representation. Despite the elimination of cheating, our approach achieves comparable quantitative results for test images, while increasing the consistency for novel view synthesis in motion. Due to its design, our hierarchical approach is only 4% slower on average than the original Gaussian Splatting. Notably, enforcing consistency enables a reduction in the number of Gaussians by approximately half with nearly identical quality and view-consistency. Consequently, rendering performance is nearly doubled, making our approach 1.6x faster than the original Gaussian Splatting, with a 50% reduction in memory requirements. Our renderer is publicly available at https://github.com/r4dl/StopThePop.

Multi-objective optimization of active control system using population guidance and modified reference-point-based NSGA-II

The optimization of the active control system is pivotal in achieving an acceptable performance level while minimizing control costs. The optimal location of actuators, efficient control force, and adequate vibration reduction are critical objectives in control system optimization. This study introduces a novel optimization algorithm, the Population-Guided and Modified Reference-Point Based Non-Dominated Sorting Genetic Algorithm-II (PMR-NSGA-II), tailored for actuator placement and tuning that incorporates input and output user preference. The analysis focuses on optimizing the vibration control of a 20-story steel frame building. The PMR-NSGA-II could reduce the computational expenses of three objectives optimization by guiding the initial population using prospective population prediction and orienting the search towards a reference point. The pre-calculated input energy distribution guides the determination of the prospective initial population. The reference point, along with the allowable drift and actuator capacity constraints, focuses the search process to efficiently obtain the Pareto fronts with less effort wasted to explore the less preferred areas in the search space. The non-dominated sorting and elitist operators are also employed to fasten the convergence. The structural analysis is conducted via non-linear time history analysis with seven ground motions considered. The PMR-NSGA-II exhibits significant computational efficiency in the active control system optimization. Results demonstrate that PMR-NSGA-II could provide the near-optimal solution closest to the reference point with a smaller population size and a faster convergence rate than the NSGA-II. The computational time per generation of the proposed PMR-NSGA-II is 1.36 to 2.43 times faster than NSGA-II for the seven ground motions analyzed. By applying PMR-NSGA-II, the number of individuals that need to be assessed to reach convergence is reduced by 50 % to 88 % compared to NSGA-II. Ultimately, the near-optimal configurations could significantly reduce the building responses and increase the performance level of the building based on FEMA 356 and ASCE 41 acceptance criteria.

Comparative Analysis of Sorting Algorithms: TimSort Python and Classical Sorting Methods

The sorted() function within the Python programming language has emerged as the primary choice among developers for sorting operations. Consequently, this study offers a comparative analysis of various classical sorting algorithms and Python's built-in sorting mechanisms, with the objective of identifying the most time-efficient sorting algorithm. The analysis involves assessing the time complexity of each algorithm while handling data arrays ranging from 10 to 1,000,000 elements using Python. These arrays are populated with randomly generated numeric values falling within the range of 1 to 1000. The benchmark algorithms utilized encompass Heap Sort, Shell Sort, Quick Sort, and Merge Sort. A looping mechanism is applied to each algorithm, and their execution speeds are gauged utilizing the Python 'time.perf_counter()' library. The findings of this study collectively indicate that Python's standard algorithm, surpasses classic sorting algorithms, including Heapsort, Shellsort, Quicksort, and Mergesort, in terms of execution.

IMPLEMENTATION OF IO-LINK TECHNOLOGY INTO THE HANDLING AND SORTING SUB-STATION OF THE FESTO FMS 500 AUTOMATED LINE

The article deals with implementing IO-LINK components in the sorting station of the FESTO FMS 500 automated line and their experimental verification based on the principles of IO-LINK technology. The implementation of such technology also includes the upgrade of the PLC control to PLC S7-1500 (SIEMENS), the addition of an IO-LINK device (BALLUFF BNI005H), and a signaling light device from the same company (SMARTLIGHT BNI007F), which ensures the indication of workplace state. Communication between the PLC device and the IO-LINK devices is carried out via the PROFINET communication network. Installing the signaling light device at the sorting station of the automated line provides light signaling messages for the operator (concerning the standard STN EN 61310-1). For example, the indication of the filling of the reserve warehouse using a light signaling message means the operator needs to empty the individual warehouses to repeat the sorting operation. In addition, part of the article is the implementation and evaluation of experimental measurements, the aim of which is to determine both the individual sorting times of various components, as well as determining their maximum number during the sorting process in connection with increasing the level of automation of the sorting station as such.

A Dual-Competition-Based Particle Swarm Optimizer for Large-Scale Optimization

Large-scale particle swarm optimization (PSO) has long been a hot topic due to the following reasons: Swarm diversity preservation is still challenging for current PSO variants for large-scale optimization problems, resulting in difficulties for PSO in balancing its exploration and exploitation. Furthermore, current PSO variants for large-scale optimization problems often introduce additional operators to improve their ability in diversity preservation, leading to increased algorithm complexity. To address these issues, this paper proposes a dual-competition-based particle update strategy (DCS), which selects the particles to be updated and corresponding exemplars with two rounds of random pairing competitions, which can straightforwardly benefit swarm diversity preservation. Furthermore, DCS confirms the primary and secondary exemplars based on the fitness sorting operation for exploitation and exploration, respectively, leading to a dual-competition-based swarm optimizer. Thanks to the proposed DCS, on the one hand, the proposed algorithm is able to protect more than half of the particles from being updated to benefit diversity preservation at the swarm level. On the other hand, DCS provides an efficient exploration and exploitation exemplar selection mechanism, which is beneficial for balancing exploration and exploitation at the particle update level. Additionally, this paper analyzes the stability conditions and computational complexity of the proposed algorithm. In the experimental section, based on seven state-of-the-art algorithms and a recently proposed large-scale benchmark suite, this paper verifies the competitiveness of the proposed algorithm in large-scale optimization problems.

Optimizing multi-spectral ore sorting incorporating wavelength selection utilizing neighborhood component analysis for effective arsenic mineral detection

Arsenic contamination not only complicates mineral processing but also poses environmental and health risks. To address these challenges, this research investigates the feasibility of utilizing Hyperspectral imaging combined with machine learning techniques for the identification of arsenic-containing minerals in copper ore samples, with a focus on practical application in sorting and processing operations. Through experimentation with various copper sulfide ores, Neighborhood Component Analysis (NCA) was employed to select essential wavelength bands from Hyperspectral data, subsequently used as inputs for machine learning algorithms to identify arsenic concentrations. Results demonstrate that by selecting a subset of informative bands using NCA, accurate mineral identification can be achieved with a significantly reduced the size of dataset, enabling efficient processing and analysis. Comparison with other wavelength selection methods highlights the superiority of NCA in optimizing classification accuracy. Specifically, the identification accuracy showed 91.9% or more when utilizing 8 or more bands selected by NCA and was comparable to hyperspectral data analysis with 204 bands. The findings suggest potential for cost-effective implementation of multispectral cameras in mineral processing operations. Future research directions include refining machine learning algorithms, exploring broader applications across diverse ore types, and integrating hyperspectral imaging with emerging sensor technologies for enhanced mineral processing capabilities.

High‐quality and efficient recovery of waste agricultural mulch film mixture via solid‐state shear milling technology

High‐quality and efficient recovery of waste agricultural mulch film mixture via solid‐state shear milling technology

Abstract 2351: Genomics annotation and interpretation in somatic oncology using structured data from a clinical knowledgebase

Abstract Accurate and comprehensive interpretation of genomic variants has become a bottleneck in clinical sequencing applications due to the accelerating precision oncology and biomedical information explosion. This motivated us to build Ephesus, a framework enabling curation of predictive, prognostic and diagnostic evidence for clinical biomarkers in cancers. Currently Ephesus is the primary content source for Roche navify Mutation Profiler (nMP). The Ephesus’ data model ensures adherence to best practices in clinical genomic content curation. Variant classification follows the AMP guidelines for somatic variant interpretation. Variant interpretations are derived from international regulatory approvals and clinical practice guidelines (FDA, EMA, TGA, eVIQ, etc.) and recommendations (NCCN, ESMO). The data model is mapped to a set of relational tables. To enforce data integrity and validity, strategies such as data normalization based on standard nomenclatural ontologies, a submit-review-approval workflow, and biological constraints are adopted. Ephesus is deployed as a web application used by curators inside and outside Roche. The UI allows users to conduct edits, filtering, sorting and bulk operations on entities by attributes. In order to minimize manual effort and maximize content coverage, inference rules are applied before ingesting into nMP. Currently Ephesus is developed for data collection, browsing and summarizing primary entities such as biomarkers, evidence items, genes, variants, variant groups, and drugs. The content from the 2023-Aug snapshot contains 11,596 directly curated biomarker profiles (including biomarker combinations), and 6M+ expanded profiles for 41 major cancer types. To evaluate the clinical reporting value of the curated knowledge, ~160k real cancer patient samples of the AACR GENIE project were queried on Ephesus. Compared with three other major knowledgebases, Ephesus/NMP is observed showing the highest performance. Table 1. Percentage of patients or biomarkers with at least one interpretation from each knowledgebase CGI (2022-10-17) CIVIC(2023-09-01) ClinVar(2023-09-09) NMP(2023-08-11) AACR-Genie-v14.0 Total patient % with any interpretation 25% (40,545) 46% (73,835) 81% (130,882) 89% (142,905) 160,965 biomarker % (including combos) with any interpretation 0.03% (256) 0.06% (526) 13.26% (122,424) 73.79% (681,128) 923,093 Citation Format: Jian Li, Lili Niu, Shuba krishna, Fnu Kinshuk, Martin Jones, Carlos Hernandez, Michael Clark, Sandra Balladares. Genomics annotation and interpretation in somatic oncology using structured data from a clinical knowledgebase [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2351.

Flexible multimode sensors based on hierarchical microstructures enable non-destructive grading of fruits in cold chain logistics

In practical applications, flexible pressure sensors must demonstrate adequate sensitivity, durability, and the ability to detect both dynamic and static forces across a wide range. The objective of this study is to develop a flexible dual-mechanism piezoelectric/piezoresistive sensor (FDMPS) based on layered microstructures to enable versatile detection in real fruit sorting operations. The FDMPS consists of an upper layer featuring planar MXene electrodes on a PDMS film, a mid-layer comprising microstructured Ag electrodes on a PVDF-TrFE/Silica gel/ZnO film, and a lower layer with microstructured MXene electrodes on a PDMS film composition. To ensure a secure fit, the three-layer structure is treated with APTES and plasma. All electrodes are produced using a pneumatic direct-write process, while the PDMS and piezoelectric films are created via a spin-coating process, making them suitable for large-scale production. The flexible FDMPS, with a 10 wt% ZnO content, achieves an optimal piezoelectric output of 3.6 V. Additionally, the FDMPS demonstrates excellent linearity (0.997), resistive sensitivity (23.65 kpa−1), and stability (5000 cycles). The incorporation of microstructures significantly enhances the performance of piezoelectric/piezoresistive sensing. Moreover, the FDMPS can accurately measure bending strain rate and angle within the ranges of 0–90°/s (with a sensitivity of 0.014 V/(°·s−1)) and 0–110° (with a sensitivity of 0.216/°), respectively. In a wireless, real-time mode, the FDMPS proves effective in monitoring the reciprocating motion of a robotic arm and assessing fruit ripeness during the grading process. This advancement promotes the development and application of precision agriculture and wearable sensing technologies.

Quantitative Assessment of Post-Harvest Losses of Guava (Psidium guajava) in Rajasthan

The present research was done for the quantitative assessment of post-harvest losses in guava. Factors responsible for the post-harvest losses in guava were identified and evaluated. The study was performed in Sawai Madhopur district of Rajasthan because the district has the highest guava production (27452 MT) among other districts of Rajasthan. The research was performed during the year of 2022. The data was taken from 40 farmers through judgmental sampling. Modified average formula was employed to quantify the post-harvest losses in guava by taking consideration of all types of losses at 5 level (collection, sorting, grading, packaging and transportation). Friedman’s test was employed to test the difference between factors responsible for the post-harvest losses in guava. The overall post-harvest losses in guava at farm level was 14.05 per cent at district level which was comparatively higher than NABCONS 2022 study on post-harvest losses. The post-harvest losses during collection & transportation to storage house (1.91%), sorting (7.50%), grading (1.73%), packaging (1.32%) and transportation (1.59%) operations at farm level majorly contributed to the post-harvest losses in guava. The highest loss (7.50%) was found during sorting operation among other operation at farm level. Friedman’s test results reflected that inappropriate sorting & grading, heavy rainfall, insect infestation, disease infection and market infrastructure were the major contributing factors for the post-harvest losses in guava.

On intelligent object sorting and assembly: versatile end-effector for robotized handling of electrical components

In the ever-evolving landscape of industrial automation, versatile and ‘smart’ robotic solutions for sorting and assembly operations have become increasingly important. The diversity of product variants in terms of size, geometry, material and texture increases the complexity of the tooling needed for robot handling during assembly. To mitigate this challenge, this study presents a novel design of a versatile and cost-effective robot end-effector, to sort out products from a stack of multi-variant objects and assemble them with minimal re-grasping adjustments. The proposed smart adaptable gripper is equipped with multiple suction heads, each with independent triggering capabilities, enabling the end-effector to adapt to diverse parts’ geometry. A vision-based system has also been integrated to the gripping system enhancing it with perception capabilities. A two-step approach is employed to address the kitting to assembly problem from start to end utilising real-time pixel-wise grasp quality analysis in combination with gripper’s masks convolution for optimized grasping. The performance of the proposed framework in terms of the design functionality is evaluated in a case study involving the manipulation and assembly of electronics on a control panel derived from the elevators’ manufacturing industry.

Research on phosphorite photoelectric sorting process based on fuzzy algorithm

Abstract In this paper, we synthesize the differences in grayscale and texture information of phosphorite X-ray images and extract a total of six feature quantities under their grayscale and texture features, respectively. After feature fusion, a fuzzy support vector machine classifier with a normal plane-type affiliation function is used to optimize the binary classification method of traditional support vector machines, which improves the accuracy of classification. To improve the sorting model’s division of the optimal classification hyperplane, the centroid position of the two types of samples is considered comprehensively. In order to give full play to the classification performance of the sorting model, the particle swarm algorithm is further used to optimize the core parameters of the fuzzy support vector machine, the penalty factor, and the kernel parameters to complete the construction of the classifier model. In the face of the mixed colloidal phosphorite that can not be exploited, based on the PSO-NP-FSVM photoelectric sorting operation, the concentrate yield rate, operation recovery rate, and optical sorting waste rate are 56.52%, 20.87%, and 71.17%, respectively. After photoelectric sorting, the grade of concentrate P2 O5 can be increased to 20.98%, which can meet the requirements for the grade of the processing plant. Therefore, the photoelectric sorting process combined with a fuzzy algorithm is capable of meeting the actual production requirements.

Microfluidic device for the high-throughput and selective encapsulation of single target cells.

Droplet-based microfluidic technologies for encapsulating single cells have rapidly evolved into powerful tools for single-cell analysis. In conventional passive single-cell encapsulation techniques, because cells arrive randomly at the droplet generation section, to encapsulate only a single cell with high precision, the average number of cells per droplet has to be decreased by reducing the average frequency at which cells arrive relative to the droplet generation rate. Therefore, the encapsulation efficiency for a given droplet generation rate is very low. Additionally, cell sorting operations are required prior to the encapsulation of target cells for specific cell type analysis. To address these challenges, we developed a cell encapsulation technology with a cell sorting function using a microfluidic chip. The microfluidic chip is equipped with an optical detection section to detect the optical information of cells and a sorting section to encapsulate cells into droplets by controlling a piezo element, enabling active encapsulation of only the single target cells. For a particle population including both targeted and non-targeted particles arriving at an average frequency of up to 6000 particles per s, with an average number of particles per droplet of 0.45, our device maintained a high purity above 97.9% for the single-target-particle droplets and achieved an outstanding throughput, encapsulating up to 2900 single target particles per s. The proposed encapsulation technology surpasses the encapsulation efficiency of conventional techniques, provides high efficiency and flexibility for single-cell research, and shows excellent potential for various applications in single-cell analysis.