Selecting Mobile Services in Cloud and Edge Environment by Moth-Flame Optimization Algorithm

Introduction

With the rapid development of mobile Internet technology, service computing in the mobile environment has become a hot research field in recent years. Specifically, with the emergence of edge computing, services can be deployed closer to mobile users to provide corresponding functionalities. Compared with traditional cloud computing, services deployed in the edge computing environment can effectively reduce the distance to users, ensure efficient network operation delivery, and realize the service interaction experience of high performance, low latency, and high bandwidth (Corcoran et al., 2016). However, due to the mobile characteristics of users and the limited resources of edge devices, the services provided by edge devices alone cannot meet a large number of increasingly complex computing requirements (Personè & Grassi, 2019). Cloud computing is good at global, non-real-time, and long-term computing, while edge computing is more suitable for local, real-time, and short-term computing. Therefore, cloud computing and edge computing can complement each other to match complex user demand scenarios and enlarge the application value of cloud computing and edge computing (Wu et al., 2019).

In reality, as users’ requirements are usually complex, a single service with limited functionality can not meet users’ expectations to use the services for complex tasks (Liu et al., 2016). Service composition combines multiple existing services in a specific logical order to complete complex tasks that a single service can not achieve. According to Gartner’s 2021 predictions, the number of services deployed in cloud and edge computing environments will grow explosively (Wang et al., 2017). A large number of services provide users with rich resources and bring new difficulties to users. One of the difficulties is how to select appropriate cloud and edge services among many candidate services to meet the complex needs of mobile users.

Usually, the above-mentioned service selection problem in the cloud and edge computing environment is affected by several factors: wireless data transmission speed, user movement, and connections between edges and clouds (Du et al., 2019). Selecting proper services out of candidate ones in such an environment can be modeled as a multi-objective optimization problem, which evolutionary algorithms can solve (Deb, 2014). This paper proposes an approach based on moth-flame optimization algorithm (MFO) and subjective and objective weighting methods to solve the service selection problem. The reason to use the Subjective and Objective weighting methods it effectively combines the subjective will of the user with the objective properties of the service. Recent studies indicate that the Moth-Flame optimization algorithm has a fast convergence speed and global search ability to produce competitive outputs in an unknown search space.

The main contributions of this paper are as follows:

• 1.

  This paper proposes a differential evolutionary moth-flame optimization algorithm (DEMF), which introduces the crossover and mutation operators of the differential evolution algorithm into the moth-flame optimization algorithm. In doing so, the early convergence and local extremum problems in the MFO could be solved, as shown by the comparative experimental results.

• 2.

  This paper proposes a subjective and objective weight calculation method to select services. Firstly, by using bias factors, users’ subjective requirements and the services’ objective performances are weighted respectively in terms of QoS attributes. Secondly, a combination method is proposed to synthesize the overall QoS values based on subjective and objective values of services. This method is used in the MFO to select a proper set of services iteratively to fulfill users’ needs.

In the rest of this paper, preliminary knowledge, the framework of proposed approach, data structure and algorithms, experimental results, related work and the conclusion are introduced and provided.

Preliminary

This section introduces the necessary knowledge of service, mobile path, mobile edge calculation formula, and the moth-flame optimization algorithm.