Table 1 Performance comparison of the proposed ensemble DL model with existing methods for IoT cyber-attack detection in smart City environments.

Chen et al¹¹

For efficient task scheduling and resource allocation for large-scale EC in smart cities.

Regional partitioning, voronoi diagrams, tetris offloading, adaptive allocation, online learning

Large-scale EC workloads

Task deadline violation rate

Ai-Quayed et al.¹²

To improve secure and efficient routing in cognitive networks for smart cities using RL and BC.

RL, BC, fault detection and validation, adaptive resource management

Simulated cognitive network data

Network efficiency, security metrics

Sahu et al.¹³

To optimize smart parking management using a multi-objective framework integrating DT and hybrid optimization models.

DT, MDP, PSO, PFO

Simulated smart parking data

Search time, energy usage, traffic congestion

Chen et al¹⁴

To enhance GSW for unified, collaborative edge-side observation and dynamic resource management.

IoT protocol integration, stm, sensorml and sos extension, dynamic task allocation

City sensing base station data

Resource accessibility, real-time processing

Wang et al.¹⁵

To reduce latency and delay in EC for smart urban vehicle transportation using hybrid optimization.

p-H-WDFOA, 5G multi-access EC, hybrid edge-cloud architecture

Smart urban vehicle network data

Latency, processing time, energy consumption

Tian et al.¹⁶

To enhance security and anomaly detection in IIoT EC for smart cities using hybrid learning models.

ELM-RNN, DRL-DQN, distributed authorization mechanism, STAPPA, GA

Simulated IIot network data

Detection rate, accuracy

Xu, Nagothu, and Chen¹⁷

To develop an autonomous and resilient EC architecture integrating AI, SDN, and BC for secure and efficient IoT ecosystems.

AR-EC, SDN, BC, LLM, federated microchain blockchain, autonomous edge resource coordination

ITS data

Network resilience, security, resource optimization

Sun et al.¹⁸

To optimize resource allocation and task scheduling in hierarchical EC networks for smart cities to ensure quality of service.

AM, priority determination, Q-Learning

Simulated edge network data

Resource utilization, task delay, QoS guarantee

Wang, Wang, and Du¹⁹

To develop an intuitive system to assist users with severe disabilities in smart home environments.

NeuroSpatialIOT, Eye Tracking, DL, context-aware control display

Eye movement and Spatial data from users

Usability score, task completion time

Far et al.²⁰

To explore the integration of BC and DRL to enhance privacy, security, and transmission efficiency in IoT-assisted smart cities.

BC, DRL, IoT system clustering and categorization, taxonomy development

Literature review (2015–2024)

Privacy, security, transmission efficiency

Khan et al.²¹

To develop an energy-efficient DL-based mechanism for optimal parallel computation offloading in mobile EC.

EPCOD, EEDOS, multi-factor analysis

Large simulated offloading data

Latency, energy consumption, accuracy

Mishra and Chaurasiya²²

To develop a hybrid DL model to secure IoT-based smart city transactions by detecting and preventing cyber-attacks.

LSTM-SVM, Min-Max normalization, weighted average filtering, RSA, BC

Smart city transaction data

Accuracy, specificity, f1 score

Ficili et al.²³

To explore and analyze innovative integration approaches to enable real-time decision-making and predictive analytics.

IoT Integration, CC, EC, AI

Case studies and applications

Latency, decision accuracy

Wang et al.²⁴

To develop an AI-based model for secure and accurate load management in IoT-based smart cities.

MSLL, MMStransformer, load forecasting, security integration

Smart city IoT data

Prediction accuracy, computational efficiency

Lilhore et al.²⁵

To develop a hybrid model for dynamic and efficient resource allocation and workload scheduling in IoT edge-cloud ecosystems.

DQN, PPO, GNN, RL

Google cluster data, alibaba cluster trace, microsoft azure traces

Scheduling time, operational cost, energy consumption

Ahmed and Elena²⁶

To investigate how integrating AI with EC enhances real-time decision-making, scalability, and network resilience in autonomous networks.

ML, FL, intelligent orchestration

Various real-world IoT and network data

Latency, scalability, resource utilization

Qasim Jebur Al-Zaidawi and Çevik²⁷

To optimize DL for real-time IoT anomaly detection with balanced scalability and efficiency using hybrid optimization.

HGWOPSO, HWCOAHHO, MCDM, AHP

Real-world IoT network data

Accuracy, precision, recall, f1-score

Kumar and Neduncheliyan²⁸

To develop an ensemble DL model for accurate IoT cyber-attack detection.

CNN, BiGRU, SSOFFN, FC

ToN-IoT dataset

Detection rate, auc, accuracy, precision, recall, f1-score