Multi-objective scheduling technique based on hybrid hitchcock bird algorithm and fuzzy signature in cloud computing
Introduction
Cloud computing is a popular computational environment that allows users to scale their resources quickly. Cloud user sends tasks while the provider of cloud presents resources for the execution of tasks. Both participants have various goals: the provider tries to obtain as many profits as possible and enhance the resource utilization while the user attempt to execute tasks with minimum execution time and cost (Mansouri and Javidi, 2020b, Ghahramani et al., 2017, Mansouri and Javidi, 2018, Mansouri et al., 2020b). In the performance of cloud computing, task scheduling plays a vital role in which resources are allocated and adjusted between different tasks. From the booming economy of information processing, designing an efficient task scheduling algorithm has attracted worldwide attention to enhance the quality of service (QoS) (Mansouri and Javidi, 2020a, Alboaneen et al., 2021). The task scheduling process in the cloud is explained in Fig. 1. Firstly, user tasks are submitted to a waiting queue and then the scheduler (i.e., data center broker) gives the necessary information (i.e., available cloud resources) from cloud information services (CIS). Then, it allocates tasks to suitable virtual machines based on the scheduling algorithm. Finally, the selected machines execute the assigned tasks by obtaining the necessary data. Generally, task scheduling algorithms are divided into two categories (i.e., static and dynamic). The static scheduling algorithm needs advanced information about tasks and environments. They are suitable for a system with less workload variation. While dynamic scheduling algorithm monitors the system continuously and can balance the workload. In general, the task scheduling problem is modeled as follows:
Assigning n tasks of users () to m heterogeneous virtual machines where n m subject to some constraints to optimize some objective functions.
Due to huge computational demands, energy consumption is one of the critical challenges in cloud data centers, which affects the benefit of service providers and meets the standard set in green computing. In Amazon, more than 50% of the budget for data center management is related to powering and refrigeration of servers. Among different ways of energy-efficient, task scheduling is at the heart of successful power reduction and resource management in cloud computing. A lot of studies have already been done to propose efficient task scheduling algorithms in recent years. In Pradhan et al. (2021), the most recent relevant literature is surveyed and the results indicated that service providers’ objectives (i.e., throughput, energy consumption, load balancing, and resource utilization) have not been considered broadly during task scheduling simultaneously. Most of them focused on execution time and did not achieve an acceptable trade-off between QoS factors and energy consumption. For example, Elaziz et al. (2019) introduced a new task scheduling strategy based on the MSA and DE algorithms to reduce makespan. Nevertheless, they did not analyze resource utilization and energy consumption. Chaudhary and Kumar (2019) presented a new scheduling technique based on Hybrid Genetic-GSA (HG-GSA) to reduce execution cost and transfer cost. But, they did not mention makespan and energy consumption.
However, heterogeneous and dynamic behavior, various Quality of Service (QoS) parameters, and unified load balancing are the difficulties of task scheduling design in cloud computing. Moreover, the task scheduling problem is one of the most common combinatorial optimization problems to optimize the key performance factors. Many previous task scheduling algorithms are limited by the fact that they assume only a small number of performance factors. In this work, we tackle the problem of developing a task scheduling strategy to optimize makespan, resource utilization, energy consumption, latency, and degree of load balance. Since low makespan and delay values show that tasks are completed within the desired time and the high degree of load balance ensures that the flexibility and scalability of data centers are well. In addition, improving the utilization of resources can solve the challenge of over-provisioning and under-provisioning and energy efficiency provides environmental and economic implications.
In this regard, different algorithms have been introduced for addressing scheduling problems (e.g., Minimum Completion Time (Madni et al., 2017), First Come First Serve (Kaur and Kinger, 2014), and Min–min (Kaur and Kaur, 2015) and Max–min (Konjaang et al., 2016)). Although these algorithms have been adopted on many cloud computing systems, some research works have verified that these strategies are inappropriate for large-scale scheduling problems (Kaur and Kaur, 2015). Since task scheduling is an NP-hard problem due to its complexity and the necessary time for finding the solution varies by the problem size (Cui et al., 2017). We know that the meta-heuristic techniques can handle a large search space of scheduling problems and determine an optimal solution in polynomial time (Ebrahimzade et al., 2020). In 2021, Houssein et al. (2021) provided a review of task scheduling based on meta-heuristics and the results represented that PSO, GA, and ACO techniques have been widely applied by researchers. Currently, most scheduling algorithms that are used to allocate tasks are not accurate and unable to satisfy the needs of cloud computing development (Alsaidy et al., 2020, Samuel Raj and Rachel, 2013) especially when the number of objectives is increased. In addition, each meta-heuristic technique has some drawbacks. For example, PSO has a slow convergence rate and weak local search. GA has a problem with complex functions in the selection and crossover operators. ACO shows complex computation and complexity in the coding scheme. The nature of most meta-heuristic techniques is such that they easily trap in local optima and hence it is necessary to combine them with suitable techniques and enhance the qualities of solutions. For instance, Madni et al. (2019) introduced a novel scheduling algorithm based on CSA to reduce execution time. But all agents of CSA search the problem space in the same way and hence premature convergence may occur. Moreover, resource utilization and energy consumption are not included.
In the same context, a new meta-heuristic technique called Hitchcock Bird Inspired Algorithm (HBIA) (Morais et al., 2019) has been presented. It is a global optimization inspired by the aggressive bird behavior portrayed by Alfred Hitchcock in the 1963 thriller “The Birds”. The HBIA algorithm is based on the attack pattern of birds and has the stages of lurking, attack, and reorganization. It is successful in solving various cost functions for a variety of domains (Morais et al., 2019, Morais et al., 2020). Therefore, we select HBIA for solving task scheduling problems and finding more satisfactory solutions in large search spaces.
Nevertheless, it suffers from several limitations including the inability to explore the search space effectively and provide a balance between exploration and exploitation activities. These limitations can be observed in its behavior during the updating process (i.e., attacking phase) and generating new solutions at the end of each iteration (i.e., reorganization phase). We know that HBIA exploration depends on replacing the dead birds (i.e., the solution with the lower quality compared to the average quality of population) with new birds. In this regard, the worst solutions obtained are replaced with randomly generated new solutions. However, this process is not efficient to handle this issue and HBIA still needs more improvements. In other words, it must consider an appropriate process which blowing fresh blood to the HBIAs’ population. As a result, we firstly propose the improved version of HFHB based on the Crow Search Algorithm (CSA) (Askarzadeh, 2016), fuzzy theory (Gupta and Saini, 2017), and Levy flight concept (Elaziz et al., 2019) to overcome the mentioned challenges. Then, the improved strategy is adapted to solve the task scheduling problem in a cloud environment. The proposed algorithm (HFHB) is similar to HBIA in its starting stage, which generates a set of solutions. But the difference is in the main loop, which consists of the modified HBIA’s attack phase and HFHB uses the CSO algorithm to improve the exploration activities. Moreover, it has a new parameter called “Attack Regulator (AR)” to balance exploration and exploitation activities with fuzzy theory. Finally, it determines the worst solutions at the end of each iteration and replaces them with the Levy flight technique.
As a brief, the main contributions of this paper can be explained as follows:
- •
Introduce HFHB as a modified version of the Hitchcock Bird Inspired Algorithm (HBIA) with the combination of the CSO algorithm to improve the exploration phase of HBIA.
- •
Apply Sugeno-signature fuzzy to provide a balancing between exploration and exploitation.
- •
Use two techniques (i.e., Levy flight and fuzzy theory) to improve the rate of convergence, the diversity of population, and the quality of solutions.
- •
Add two methods (i.e., crowding distance and ranking non-dominated solution) into HFHB and introduce the multi-objective form of HFHB that is named MOHFHB.
- •
Model task scheduling as a multi-objective optimization problem and solve it based on fuzzy theory to optimize the key performance indicator parameters.
- •
Try to balance the trade-off among the five optimization objectives (i.e., makespan, resource utilization, energy consumption, latency, and degree load balance) during task scheduling.
- •
Present the comparative and statistical analysis of HFHB by 20 test functions to ensure its capability in solving the global optimization problems compared with Multi-Verse Optimizer (MVO), Crow Search Algorithm (CSA), Hitchcock Bird inspired Algorithm (HBIA), Moth–Flame Optimizer (MFO), Archimedes optimization algorithm (AOA), Arithmetic Optimization Algorithm (ARO), and Grey Wolf Optimizer (GWO).
- •
Provide the comparative and statistical analysis of MOHFHB by ZDT test functions to ensure its capability in solving the multi-objective problems compared with Multi-objective Multi-Verse Optimizer (MOMVO), Multi-objective Salp Swarm Algorithm (MSSA), and Non-dominated Sorting Genetic Algorithm (NSGA-II).
- •
Evaluate the performance of HFHB in terms of makespan, resource utilization, energy consumption, latency, degree imbalance, throughput as a task scheduler with the different number of tasks. In addition, evaluate MOHFHB in terms of energy consumption, degree imbalance, and makespan.
The rest of paper is divided into the following sections: Section 2 explains a brief discussion of the related works on task scheduling. Section 3 illustrates the necessary backgrounds about the fuzzy signature, Crow Search Algorithm (CSA), multi-objective problems, and Hitchcock Bird-Inspired Algorithm (HBIA). Section 4 brings our proposed approach. Section 5 models the task scheduling problem and explains how to solve it with the proposed algorithm Section 6 discusses the evaluation results and the statistical test. Section 7 presents our conclusion and future works.
Section snippets
Related works
Meta-heuristics are preferred in different commercial and industrial applications since they are derivative-free and can solve large problems in a suitable time (Panagant et al., 2019, shafaei and Khayati, 2020, Dehestani et al., 2021, Meng et al., 2021). According to the literature, the optimization problem can be single-objective and multi-objective. When there are more than three objective functions, determining the optimal solution is difficult and hence meta-heuristic techniques present
Preliminaries
This section introduces the basic concepts of the fuzzy signature, Hitchcock bird-inspired algorithm (HBIA), and the Crow Search Algorithm (CSA).
The proposed algorithm (Hybrid Fuzzy Hitchcock Bird algorithm)
This section provides the general idea and development of a novel Hybrid Fuzzy Hitchcock bird (HFHB) algorithm and its multi-objective form. Typically, the success of any optimization algorithm depends on considering trade-off in exploration (i.e., global search) and exploitation (i.e., local search) capabilities, providing diversity in population, and initializing appropriate individuals.
HBIA attempts to initialize the first population so that the population is filled with appropriate
Basic descriptions
A large number of users executed their tasks in the cloud environment and so task scheduling plays a vital role in resource utilization, turnaround time, latency, and load balancing. The task scheduling problem can be presented as follows. The inputs are:
- •
A set of n tasks that is defined as where each task has a particular size (TS) and a deadline (D) for execution.
- •
A set of resources that is defined as where each resource has several features (i.e., CPU speed (CS), RAM
Experiments and discussion
In this section, the performance of the proposed algorithms (i.e., HFHB and MOHFHB) are evaluated in two parts. In the first part, they are evaluated as two global optimization techniques. We use two types of datasets that one dataset is appropriate for evaluating the performance of single-objective methods and the other one is appropriate for evaluating the performance of multi-objective methods.
In the second part, the proposed algorithms (i.e., HFHB and MOHFHB) are evaluated as two methods
Conclusion
In the cloud environment, task scheduling is one of the fundamental problems and many algorithms have been presented to solve this issue. In this paper, two versions of an algorithm to solve scheduling problems based on Hitchcock Bird Inspired Algorithm (HBIA) and Crow Search Algorithm (CSA) are proposed. In HFHB as single-objective version, five conflicting objectives namely makespan, resource utilization, energy consumption, latency, and degree load balance are considered and solutions are
CRediT authorship contribution statement
B. Mohammad Hasani Zade: Methodology, Software (Programming and development), Writing - original draft, Writing - review & editing. N. Mansouri: Software (Programming and development), Formal analysis, Writing - original draft, Writing - review & editing. M.M. Javidi: Conceptualization (Ideas), Investigation, Visualization.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (104)
- et al.
A metaheuristic method for joint task scheduling and virtual machine placement in cloud data centers
Future Gener. Comput. Syst.
(2021) A novel metaheuristic method for solving constrained engineering optimization problems: crow search algorithm
Comput. Struct.
(2016)- et al.
Enhanced particle swarm optimization for task scheduling in cloud computing environments
Procedia Comput. Sci.
(2015) - et al.
On convergence of the multi-objective particle swarm optimizers
Inform. Sci.
(2011) - et al.
Task scheduling in cloud computing based on hybrid moth search algorithm and differential evolution
Knowl.-Based Syst.
(2019) - et al.
A novel fuzzy-PSO term weighting automatic query expansion approach using combined semantic filtering
Knowl.-Based Syst.
(2017) - et al.
MORM: A multi-objective optimized replication management strategy for cloud storage cluster
J. Syst. Archit.
(2014) - et al.
Hybrid task scheduling strategy for cloud computing by modified particle swarm optimization and fuzzy theory
Comput. Ind. Eng.
(2019) Moth-flame optimization algorithm: A novel nature-inspired heuristic paradigm
Knowl.-Based Syst.
(2015)- et al.
Salp swarm algorithm: A bio-inspired optimizer for engineering design problems
Adv. Eng. Softw.
(2017)
Optimization of problems with multiple objectives using the multi-verse optimization algorithm
Knowl.-Based Syst.
The whale optimization algorithm
Adv. Eng. Softw.
Grey wolf optimizer
Adv. Eng. Softw.
Task scheduling in heterogeneous cloud environment using mean grey wolf optimization algorithm
ICT Express
Implementing of the multi-objective particle swarm optimizer and fuzzy decision-maker in exergetic, exergoeconomic and environmental optimization of a benchmark cogeneration system
Energy
Butterfly optimization algorithm for optimum shape design of automobile suspension components
Mater. Test.
The arithmetic optimization algorithm
Comput. Methods Appl. Mech. Engrg.
Multi-verse optimizer: theory, literature review, and application in data clustering
Nat.-Inspired Optim.
Heuristic initialization of PSO task scheduling algorithm in cloud computing
J. King Saud Univ. - Comput. Inf. Sci.
Artificial flora optimization algorithm for task scheduling in cloud computing environment
Robot cooperation without explicit communication by fuzzy signatures and decision trees
Scheduling semiconductor testing facility by using cuckoo search algorithm with reinforcement learning and surrogate modeling
IEEE Trans. Autom. Sci. Eng.
Self-adaptive many-objective meta-heuristic based on decomposition for many-objective conceptual design of a fixed wing unmanned aerial vehicle
Aerosp. Sci. Technol.
Cost optimized hybrid genetic-gravitational search algorithm for load scheduling in cloud computing
Appl. Soft Comput.
Levy flights in metaheuristics optimization algorithms–a review
Appl. Artif. Intell.
Cloud service reliability modelling and optimal task scheduling
IET Commun.
An improved optimization model to predict the microhardness of Ni/Al2o3 nanocomposite coatings prepared by electrodeposition: A hybrid artificial neural network-modified particle swarm optimization approach
Measurement
A new hybrid approach for reliability-based design optimization of structural components
Mater. Test.
Emosoa: a new evolutionary multi-objective seagull optimization algorithm for global optimization
Int. J. Mach. Learn. Cybern.
A PSO-based task scheduling algorithm improved using a load-balancing technique for the cloud computing environment
Concurr. Comput.: Pract. Exper.
PSO–ANN-Based prediction of cobalt leaching rate from waste lithium-ion batteries
J. Mater. Cycles Waste Manag.
An improved henry gas solubility optimization algorithm for task scheduling in cloud computing
Artif. Intell. Rev.
Scheduling dual-objective stochastic hybrid flow shop with deteriorating jobs via bi- population evolutionary algorithm
IEEE Trans. Syst. Man Cybern.: Syst.
Multiverse optimization algorithm for stochastic biobjective disassembly sequence planning subject to operation failures
IEEE Trans. Syst. Man Cybern.: Syst.
A review on swarm intelligence and evolutionary algorithms for solving flexible job shop scheduling problems
IEEE/CAA J. Autom. Sin.
Flexible job-shop rescheduling for new job insertion by using discrete jaya algorithm
IEEE Trans. Cybern.
Toward cloud computing qos architecture: Analysis of cloud systems and cloud services
IEEE/CAA J. Autom. Sin.
A guided population archive whale optimization algorithm for solving multiobjective optimization problems
Expert Syst. Appl.
Lexicographic multiobjective scatter search for the optimization of sequence-dependent selective disassembly subject to multiresource constraints
IEEE Trans. Cybern.
An efficient equilibrium optimizer with mutation strategy for numerical optimization
Appl. Soft Comput.
Fuzzy multi-objective programming problem for revenue management in food industry
J. Revenue Pricing Manag.
Archimedes optimization algorithm: a new metaheuristic algorithm for solving optimization problems
Appl. Intell.
Task scheduling in cloud computing based on meta-heuristics: review, taxonomy, open challenges, and future trends
Swarm Evol. Comput.
On the exploration and exploitation in popular swarm-based metaheuristic algorithms
Neural Comput. Appl.
Improved max–min scheduling algorithm
OSR J. Comput. Eng. (IOSR-JCE)
Analysis of job scheduling algorithms in cloud computing
Int. J. Comput. Trends Technol. (IJCTT)
Different task scheduling algorithms in cloud computing
Int. J. Latest Trends Eng. Technol.
Multi-objective optimization approach to enhance the stencil printing quality
Procedia Manuf.
Makespan improvement of PSO-based dynamic scheduling in cloud environment
Cited by (16)
E-AVOA-TS: Enhanced African vultures optimization algorithm-based task scheduling strategy for fog–cloud computing
2023, Sustainable Computing: Informatics and SystemsTask scheduling in fog environment — Challenges, tools & methodologies: A review
2023, Computer Science ReviewAn improved brain storm optimization algorithm with new solution generation strategies for classification
2022, Engineering Applications of Artificial IntelligenceCitation Excerpt :An increasing number of swarm intelligence algorithms are now being designed to find global solutions for optimization problems. Compared with traditional optimization methods, swarm intelligence algorithms have enormous advantages and are widely used to solve practical problems (Zamani et al., 2021; Zade et al., 2021; Xiao et al., 2021). Cervantes et al. (2017) introduced particle swarm optimization (PSO) algorithm into support vector machine classifier, which combining optimization techniques with classification techniques to achieve better classification performance.
Fuzzy-GEC an Energy-Aware Hybrid Task Scheduling on the Cloud
2024, Lecture Notes in Networks and Systems