Table 20.

Categorizing the state-of-the-art in other areas of applications.

Study	Attributes
Al-Qerem et al. [46]	Main concept: A new variation in the optimistic concurrency control protocol is reintroduced.
	Case study: -
	App domain: Cloud server and a fog node.
	Advantage: Reduce loss rates, restart rates, and connection delays.
	Disadvantage: The present work is only related to readable transactions and updates.
	Simulation/Implementation: Use of the iFogSim simulation tool.
	Dataset: -
	Future work: The ineffectiveness of the validation process. This is performed on the server using standard protocols. Many unnecessary transaction restarts and aborts will occur in the cloud.
Pendit et al. [16]	Main concept: Make a task-scheduling scheme that is based on RL.
	Case study: -
	App domain: Neural networks.
	Advantage: Efficient resource usage and task execution time minimization, as well as a significant reduction in communication expenses.
	Disadvantage: -
	Simulation/Implementation: Experiment with different lengths of instructions (1000–10,000).
	Dataset: -
	Future work: It is necessary to investigate the impact of various RL algorithms for scheduling tasks in the IoT context, such as prioritized deep Q-network (DQN), double DQN, and others.
Potu et al. [47]	Main concept: In fog computing, an Extended Particle Swarm Optimization (EPSO) method was developed to enhance resource efficiency and reduce the time spent completing tasks.
	Case study: Fog computing
	App domain: Cloud–fog environment
	Advantage: Improve resource efficiency and minimize time spent completing tasks.
	Disadvantage: While doing IoT of work in fog nodes helps minimize latency, it can also lead to greater fog node energy usage.
	Simulation/Implementation: iFogSim and Eclipse editor were used to implement Java.
	Dataset: From the 200 to 500 tasks, l7 datasets were created.
	Future work: Design a strategy for exchanging latency and energy consumption using parallel metaheuristic algorithms.
Kandan et al. [48]	Main concept: Designing a new optimization-based job scheduling technique for resource management and allocation in a cloud computing environment.
	Case study: -
	App domain: IoT cloud environment.
	Advantage: Minimum flow time, minimal lateness.
	Disadvantage: -
	Simulation/Implementation: Simulates Aquila’s behaviors.
	Dataset: -
	Future work: To properly schedule jobs in the IoT cloud, a combination of these two metaheuristic algorithms and deep learning algorithms might be offered.
Stavrinides et al. [18]	Main concept: Introduce a dynamic scheduling algorithm using partial calculations to use scheduling gaps to achieve real-time scheduling.
	Case study: -
	App domain: Fog computing environment.
	Advantage: Take into account the consequences of global error propagation between process component activities.
	Disadvantage: Overhead not checked.
	Simulation/Implementation: Simulation program in C++.
	Dataset: -
	Future work: The proposed scheduling technique is implemented in a three-tier environment. In circumstances of high workload, more cloud resources are used to schedule IoT workflows.
Sheng et al. [49]	Main concept: Provide a policy-based REINFORCE algorithm for scheduling tasks based on tasks and heterogeneity of resources in the EC.
	Case study: -
	App domain: Edge computing.
	Advantage: Optimize the order of work execution and job allocation jointly.
	Disadvantage: -
	Simulation/Implementation: Simulation in Python 3.
	Dataset: -
	Future work: Common cloud computing and edge computing taking into account communication latency.
Attiya et al. [50]	Main concept: To discover the best strategy for scheduling IoT applications, the ChOA was combined with the MPA and the Disruptor Operator.
	Case study: There are 1000 tasks in the simulated workload, ranging in length from 1000 to 20,000 MI. The actual workload is made up of 1000 jobs ranging in duration from 1000 to 20,000 MI.
	App domain: Fog computing.
	Advantage: Increasing fog processing power and optimizing the scheduling of IoT application tasks.
	Disadvantage: -
	Simulation/Implementation: Implemented using CloudSim
	Dataset: The “Parallel workload Archives” contains authentic datasets created by NASA Ames iPCS/860 and HPC2N.
	Future work: In a real fog computing system, implement the CHMPAD scheduling algorithm. It can also be used to handle more difficult optimization problems, adjust quadratic characteristics, vehicle routing, assembly line tuning, and healthcare planning, as well as to make other enhancements and advancements.