Multiobjective-based constraint-handling techniques are popular in evolutionary constrained single-objective optimization. However, most of these run into troubles when dealing with multiobjective optimization problems (CMOPs). That is, they have difficulty optimizing too many objective functions, ineffective maintaining population diversity, or challenged establishing appropriate additional functions. As a remedy to limitations, we propose novel technique called NRC for handling CMOPs. The...

Constraints may scatter the Pareto optimal solutions of a constrained multiobjective optimization problem (CMOP) into multiple feasible regions. To avoid getting trapped in local regions or part global regions, evolutionary algorithm (CMOEA) should consider both escape force and expansion carefully during search process. However, most CMOEAs fail to provide these two forces effectively. As remedy for this limitation, article proposes method called TPEA. TPEA maintains three populations,...

Vision-based semantic scene completion task aims to predict dense geometric and 3D representations from 2D images. However, modeling a single view is an ill-posed problem, limited by the field of occlusion problems caused image input. Moreover, existing methods tend produce erroneous hallucinations overly smooth boundary segmentation due lack information. To address this we propose MixSSC, which mixes sparsity forward projection with denseness depth-prior backward projection. The aim use...

3D Semantic Scene Completion (SSC) provides comprehensive scene geometry and semantics for autonomous driving perception, which is crucial enabling accurate reliable decision-making. However, existing SSC methods are limited to capturing sparse information from the current frame or naively stacking multi-frame temporal features, thereby failing acquire effective context. These approaches ignore critical motion dynamics struggle achieve consistency. To address above challenges, we propose a...

Camera-based 3D semantic scene completion (SSC) provides dense geometric and perception for autonomous driving. However, images provide limited information making the model susceptible to ambiguity caused by occlusion perspective distortion. Existing methods often lack explicit modeling between objects, limiting their of context. To address these challenges, we propose a novel method VLScene: Vision-Language Guidance Distillation Semantic Scene Completion. The key insight is use...

The construction of undetectable adversarial examples with few perturbances remains a difficult problem in attacks. At present, most solutions use the standard gradient optimization algorithm to build by applying global perturbations benign samples and then launch attacks on targets (e.g., face recognition systems). However, when perturbance size is limited, performance these approaches suffers substantially. content crucial places an image, other hand, will impact final prediction; if areas...

The goal of constrained multiobjective evolutionary optimization is to obtain a set well-converged and well-distributed feasible solutions. To achieve this goal, delicate tradeoff must be struck among feasibility, diversity, convergence. However, balancing these three elements simultaneously through single model nontrivial, mainly because the significance each element varies in different phases. As an alternative approach, we adapt distinct models various phases introduce novel algorithm...

The rapid growth of vehicles as countries become more developed has brought great challenges to traffic prediction. Recent works model only local or global spatial-temporal features via graph neural networks (GNNs). Furthermore, the explicit structure information may contain bias, in particular, lack connections among multiple nodes when fact, they are interdependent. This results inability accommodate interaction and underutilization high-quality information. In this article, we design an...

The increasing integration of renewable energy sources and the growing complexity modern power grids demand innovative solutions for efficient management. This paper introduces a novel dynamic demand-aware Power Grid Intelligent Pricing (PGIP) algorithm based on Deep Reinforcement Learning (DRL). proposed PGIP aims to optimize consumption pricing in real time by leveraging capabilities DRL adapt patterns evolving grid conditions. employs sophisticated neural network architecture model...

Finite-impulse response (FIR) Filter is widely used in wireless sensor networks as a signal pre-processing step. Because nodes require long working periods and ultra-low cost, traditional FIR structures are inapplicable multipliers occupy too much die size for such node's chips. This paper proposes novel filter the design of application specific integrated circuits (ASICs) nodes, which can reduce hardware cost to minimum. The experiments show that proposed structure lead significant savings...

The cloud data center provide powerful computing environment, but in practice the average utilization of servers is very low. problem how to schedule resources when loads virtual machines are heavy. In this paper, a dynamic forecast algorithm introduced. management algorithm(VM-DFM) reduces amount migration.

Multimodal Sentiment Analysis (MSA) technology, prevalent in consumer applications and mobile edge computing (MEC), enables sentiment examination through user data collected by smart devices. Despite the focus on representation learning MSA, current methods often prioritize recognition performance modality interaction fusion. However, they struggle to capture multi-view cues across different states, limiting multimodal representations' expressiveness. This paper develops an innovative MSA...

The amount of scientific data is currently growing at an unprecedented pace, with tensors being a common form that display high-order, high-dimensional, and sparse features. While tensor-based analysis methods are effective, the vast increase in size has made processing original tensor infeasible. Tensor decomposition offers solution by decomposing into multiple low-rank matrices or can be efficiently utilized methods. One such algorithm Tucker decomposition, which decomposes N -order factor...

Monocular depth estimation has made considerable advances under clear weather conditions. However, how to learn accurate scene rain and fog conditions alleviate the negative influence of occlusion, light, visibility, etc., is an open problem. To address this problem, in article, we split adverse network into two subbranches: prediction branch masked encoding branch. The used for estimation. branch, inspired by image modeling, uses random masks simulate occlusion or low visibility often seen...

The non-technical losses caused by abnormal power consumption behavior of users seriously affect the revenue companies and quality supply. To assist electric in improving efficiency audit regulating users, this paper proposes a anomaly detection method named High-LowDAAE (Autoencoder model for dual adversarial training high low-level temporal features). adds an extra “discriminator” AE3 to USAD (UnSupervised Anomaly Detection on Multivariate Time Series), which performs same function as AE2...