Article Preview
Top1. Introduction And Background
SCM involves the management of flow of goods and services (Syed and Zhang, 2019). It covers the management of work-in-progress inventory, storage of raw materials and resources from point of origin to point of consumption. Such management spans manufacturing companies, distribution firms, Cloud Service Providers (CSP) etc. CSP’s nowadays make use of consolidated and large Physical Internet (PI) with active hyperconnected supply chains for the provisioning of resources. As such, quite a number of challenges are encountered by the SCM systems as mentioned in (Ray et al., 2016) to include variety, velocity, volume value and verification. Velocity is the distance covered in specific direction (Rowland, 2019). The velocity in achieving an objective is equivalent to its speed. In terms of Cloud computing, it is the time covered for the optimal allocation of resources requested by user jobs. With the hyperconnected PI, it is the allocation of Virtual Machine (VM) in form of task request to satisfy user jobs (Aliyu et al. 2019). Several models are proposed by many researchers (Maryam et al., 2019; Sruthi et al., 2019; Ibrahim et al., 2016) for task scheduling. These authors considered different factors like Internet of Things (IoT), hyperconnected logistics, service level agreement aware application deployment etc. This research work proposes to involve two search characteristics to obtain faster convergence chain in Cloud task scheduling.
Task scheduling classification according to Aliyu et al, (2019) can be categorized either as real time, cloud service, workflow, or dynamic and static scheduling. Several dynamic techniques have been reviewed based on deterministic and metaheuristics to map out task scheduling disputes. Such researches can be seen in (Najme et al., 2019; Shabeera et al., 2016) and lots more. Dhinesh and Venkata (2013) adopted Linear Programming (LP) for Facility Location Problem (FLP) to reduce cost in setting up data centers. The LP technique used follows a stringent deterministic procedure that doesn’t scale well with dynamic environments like cloud. With heterogeneous Cloud environment, deterministic techniques perform poorly in terms of optimizing supply chain in task scheduling. Metaheuristic are sets of algorithmic concepts designed to find, generate, or select a heuristic that can render a sufficiently good result to an optimization problem with little alterations. They provide good quality result through exploration and exploitation especially with limited computation capacity or incomplete information applicable to a wide set of problems. Examples of metaheuristics include, clonal selection, ACO, gentic algorithm and so on. They perform reasonably in optimizing CSP throughput in areas such as, makespan minimization, load balancing and scalability by rendering close optimal results within reasonable time period (Aliyu, et al., 2019).
CSP’s can be categorized into three basic tiers for the provisioning of cloud resources. The 1st tier actually controls the backbone which they serve as the manufacturers of all cloud services. Tier 2 purchase services from tier 1, as such becomes the retailers. Tier 3 purchase services from tier 2, as such becomes the suppliers to end users. Despite the categorization, Cloud’s efficiency needs to be enhanced and improved for faster supply chain of service distribution. In view of the above challenge, the key objective of this research work is to involve the optimization mechanism used by metaheuristic and Deterministic Algorithm (DA) so as to improve SCM in Cloud environment. It is believed that using two search characteristics by both algorithms will strengthen their behavior to promote faster convergence speed and foster SCM throughput.
Section 2 of this paper discusses the concept of ACO and formulates the supply chain system where an architectural framework for the two search characteristics was structured. The experimental setup was highlighted in section 3, where experimental results, analysis and Comparison were discussed in section 4. Related work and discussion was conducted in section 5, as section 6 concludes the entire work and provided future research study.