Received: 8 October 2018 Revised: 1 June 2019 Accepted: 1 June 2019
DOI: 10.1002/cpe.5442
RESEARCH ARTICLE
Task scheduling in Internet of Things cloud environment using
a robust particle swarm optimization
Mohammed Zaki Hasan
1,2
Hussain Al-Rizzo
2
1
College of Computer Sciences and
Mathematics, University of Mosul, Mosul, Iraq
2
Systems Engineering Department,
George W. Donaghey College of Engineering
and Information Technology, University of
Arkansas Little Rock, Little Rock, Arkansas
Correspondence
Mohammed Zaki Hasan, College of Computer
Sciences and Mathematics, University of
Mosul, Mosul 41002, Iraq; or
Systems Engineering Department,
George W. Donaghey College of Engineering
and Information Technology, University of
Arkansas Little Rock, Little Rock, Arkansas.
Email: mzallayla@ualr.edu
Summary
Internet of Things (IoT) is steadily growing in support of current and projected real-time dis-
tributed Internet applications in civilian and military applications, while Cloud Computing has
the ability to meet the performance expectations of these applications. In this paper, we present
the implementation of logistics management applications relying on cooperative resources with
optimized performances. To dynamically incorporate smart manufacturing objects into logistics
management IoT applications within a ubiquitous environment, task scheduling must be pro-
vided for resource allocation in an optimized way. Within such environment, we propose a task
scheduling algorithm based on a robust Canonical Particle Swarm Optimization (CPSO) algorithm
to solve the problem of resource allocation and management in both homogeneous and hetero-
geneous IoT Cloud Computing. Our objective is to satisfy the Makespan by performing optimal
task scheduling while considering different policies of incoming tasks. Performance evaluation
from simulation experiments reveals that optimizing the Makespan can be significantly improved
by Longest Processing Time (LPT), Shortest Processing Time (SPT), Earliest Computation Time
(ECT), Earliest Starting Time (EST), Earliest Deadline First (EDF), and Earliest Duedate (EDD)
using our CPSO algorithm as compared with traditional list task scheduling algorithms.
KEYWORDS
cloud computing, Internet of Things, quality of services, robust optimization, task scheduling
1 INTRODUCTION
Newfangled devices, systems, applications, and interfaces are nearly expected to become interconnected objects in Internet of Things (IoT).
1
IoT
allows achieving multiple tasks and glean, useful, and actionable information.
2
The IoT has witnessed explosive growth of objects driven by a
gadget of consumers to access multimedia files to accomplish several tasks that work individually or independently.
3
The underlying platform
of IoT permits objects to collaborate and connect to Internet in a smart way without human intervention (ie, smart objects).
4
Smart objects are
able to sense, interact, and create a ubiquitous environment. Within such environment, data could be collected and used to support further
decision-making such as logistic and scheduling in transportation, manufacturing, healthcare, industrial automation, and emergency handling
applications.
4
Moreover, IoT can be seen as a sort of large-scale distributed system such as in Cloud Computing where these objects have a high
degree of heterogeneity with respect to hardware and software. However, how many of them can work simultaneously and consistently? Despite
the existing connectivity solutions, more comprehensive effort is needed in the design of a new architecture IoT reference model. Consequently,
new model provision is needed for communication and resource allocation to successfully support massive IoT deployment in terms of services
and devices. The new IoT reference architecture model should scale operationally and economically with the expansion of IoT components and
provide smart functionalities for autonomous reasoning among objects.
4
Implementing IoT systems in the market will offer great opportunities and enable a variety of new devices and services to stream, connect,
and interact in single home or small business. The IoT can be viewed as large-scale distributed system where the components have a high degree
of heterogeneity with respect to both software and hardware in terms of services and devices. The Quality of Services (QoS) for users accessing
IoT supported with Cloud Computing is increasing exponentially.
3
Furthermore, the execution context is extremely dynamic, which leads to
countless, unpredictable service demands and resources that represent a complicated issue for users who do not own sufficient computing
Concurrency Computat Pract Exper. 2019;e5442. wileyonlinelibrary.com/journal/cpe © 2019 John Wiley & Sons, Ltd. 1 of 17
https://doi.org/10.1002/cpe.5442