查看更多>>摘要：A Robotic Mobile Fulfillment System (RMFS) is a parts-to-picker system designed for e-commerce warehousing where robots are used to fetch inventory pods from the storage area and transport them to the appropriate workstation. At these stations human workers pick the required amount of goods from the pods to fulfill the active orders. The RMFS is composed of several hard sub-problems which are typically solved sequentially. To the best of the authors’ knowledge, there exists no algorithm that integrates multiple of these problems and consider the interdependencies between them. This paper focuses on solving the integrated order to workstation and robot scheduling problem and proposes a bi-level memetic algorithm. Computational experiments on a wide range of problem instances show the importance of considering an integrated solution approach over a sequential approach for this complex problem. The experiment clearly show the impact of the existing interdependencies. Moreover, the study shows that the pod selection problem for the order fulfillment has a significant impact on the overall system performance. The inventory pod's consolidation opportunity and distance from the picking station has to be taken into account.
查看更多>>摘要：Biogeography-based optimization (BBO) is a swarm intelligence optimization algorithm based on migration and mutation operations, which is usually used to solve the complex optimization problems. However, it is also a challenging task for conventional BBO to solve some complex and diversified optimization problems with a perfect balance between the performance of exploration and exploitation. In this paper, we propose a variant of BBO approach called BBO with improved migration and adaptive mutation (BBOIMAM) to improve the performance of conventional BBO for dealing with different optimization problems. First, BBOIMAM introduces an improved migration strategy which includes the generalized sinusoidal migration model and immigration strategy based on elite-learning mechanism for the improvement of the local search ability. Second, we propose an adaptive mutation strategy based on the spring vibration to further enhance the population diversity, so that improving the global search ability of the algorithm. By the combination of the proposed improved migration and adaptive mutation strategies, the exploration and exploitation performance of the algorithm can be balanced. We make a large number of experiments on a set of various kinds of benchmark functions, and the experimental results demonstrate that the proposed BBOIMAM approach achieves better performance than several state-of-the-art peer algorithms on CEC 2017 and CEC 2020 test function sets and three cases of the antenna array beam pattern optimization problems.
查看更多>>摘要：Image denoising is the key component in several computer vision and image processing operations due to unavoidable noise in the image generation process. For medical image processing, deep convolutional neural networks (CNN) gives a state-of-the-art performance. However, network structures are manually constructed for specific tasks and require several trials to tune a large number of hyperparameters, which can take a long time to construct a network. Additionally, the fittest hyperparameters which may be suitable for source data properties like noisy features cannot be easily found to target data. The realistic noise is generally mixed, complex, and unpredictable in medical images, which makes it difficult to design an efficient denoising network. We developed a Differential Evolution (DE) based automatic network evolution model in this paper to optimize the network architectures and hyperparameters by exploring the fittest parameters. Furthermore, we adopted a transfer learning technique to accelerate the training process. The proposed evolutionary algorithm is flexible and finds optimistic network architectures using well-known methods including residual and dense blocks. Finally, the proposed model was evaluated on four different medical image datasets. The obtained results at different noise levels show the potentiality of the proposed model named DEvoNet for identifying the optimal parameters to develop a high-performance denoising network structure.
查看更多>>摘要：Consensus is an important issue in group decision making as it aims to avoid future contestation by the decision makers. The elicitation of criteria and decision makers’ weights is an essential part of decision-making problems. This study proposes a new consensus model for group decision making based on dual hesitant fuzzy and evolutionary algorithm for the definition of unknown criteria and decision makers’ weights. The Dual Hesitant Fuzzy Preference Relations combines the advantages of intuitionistic and hesitant fuzzy representations, and it is used to deal with the imprecision in decision makers’ judgments. In order to find decision makers weights to reach a better level of consensus without the need to modify initial assessments, the proposed Genetic Algorithm (GA) is applied. An illustrative application case is presented in a large steel company considering sustainable criteria. An instance generator was proposed to create different scenarios of decision making varying the number of criteria, decision makers, and level of hesitation. Several instances were used in the computational tests for the evaluation of the GA performance. The effectiveness of the proposed GA was verified by comparing its results with the results of the implemented Particle Swarm Optimization (PSO) algorithm. The GA yielded solutions with improved consensus levels in a reasonable runtime, especially for a small or medium number of decision makers.
查看更多>>摘要：Underexposure regions are vital in constructing a complete perception of the surrounding environment for safe autonomous driving. The availability of thermal cameras has provided an essential alternative to explore regions where other optical sensors lack in capturing interpretable signals. A thermal camera captures an image using the heat difference emitted by objects in the infrared spectrum, and object detection in thermal images becomes effective for autonomous driving in challenging conditions. Although object detection in the visible spectrum domain has matured, thermal object detection lacks effectiveness. A significant challenge is the scarcity of labeled data for the thermal domain, which is essential for SOTA artificial intelligence techniques. This work proposes a domain adaptation framework that employs a style transfer technique for transfer learning from visible spectrum images to thermal images. The framework uses a generative adversarial network (GAN) to transfer the low-level features from the visible spectrum domain to the thermal domain through style consistency. The efficacy of the proposed object detection method in thermal images is evident from the improved results when using styled images from publicly available thermal image datasets (FLIR ADAS and KAIST Multi-Spectral).
查看更多>>摘要：This study introduces a bi-objective integrated supply chain (SC) scheduling (SCS) model to deal with the challenges of providing highly customized and on-time delivery requirements at the least cost. To address these real-life challenges, the model integrates the supply portfolio into production scheduling with a customer-imposed delivery time window. A set of conflicting factors is considered for the SC in which the manufacturer is modeled using the flexible job shop (FJS) problem to include greater flexibility in process routing. Since the proposed SCS model extends the strongly NP-hard FJS problem, two new meta-heuristic algorithms are developed, enhancing the performance of the multi-objective particle swarm optimization (MOPSO). A Tabu search inspired search mechanism and a problem-specific mutation operator are designed in both algorithms. As inadequate archive diversity leads to premature convergence and unsatisfactory Pareto solutions in many cases, this work employs two practical approaches separately (i.e., crowding distance (CD) and niche count preservation of reference point (RP)) in the proposed MOPSOs, called CD-MOPSO and RP-MOPSO, respectively. For leader selection, CD-MOPSO uses the swarm distance-based Roulette wheel, whereas the concept of RP is proposed in RP-MOPSO to expedite convergence. The performances of the proposed algorithms are validated against four existing algorithms and assessed by utilizing several criteria after solving 45 artificial instances. The simulation results and statistical analysis demonstrate the supremacy of the RP-MOPSO which increases the flexibility for a decision-maker in providing a higher number of Pareto solutions and more diverse and regular frontiers within reasonable computational time. Finally, a sensitivity analysis of the model is conducted, and managerial insights are provided.
查看更多>>摘要：Incremental feature selection is an efficient paradigm that updates an optimal feature subset from added-in data without forgetting the previously learned knowledge. Most existing studies of rough set-based incremental feature selection require scanning all added-in samples and all possible candidate features when determining a best feature. However, such a classical search strategy has to perform some redundant calculations, which increase the computing and memory space resources. To avoid the redundant calculations, we propose a novel incremental feature selection method using sample selection and feature-based accelerator. First, a feature selection framework based on discernibility score is proposed as basis for our incremental method. Second, sample selection scheme is proposed to eliminate useless samples from added-in data. This scheme ensures that only useful samples are considered in the incremental process. Third, feature-based accelerator is designed to incrementally select a best feature and simultaneously remove redundant candidate features. It is theoretically guaranteed redundant features removed earlier remain redundant and will not be reexamined during the rest of the process. Finally, our incremental feature selection algorithm is designed by a two-stage procedure including sample selection scheme and feature-based accelerator. The results of experiments validate the time efficiency of the proposed incremental algorithm, especially on datasets with numerous instances or high dimensions.
NSTL
Elsevier