Elsevier

Information Sciences

Volume 182, Issue 1, 1 January 2012, Pages 77-92
Information Sciences

A self-optimizing mobile network: Auto-tuning the network with firefly-synchronized agents

https://doi.org/10.1016/j.ins.2010.11.017Get rights and content

Abstract

The reduction of cost and complexity is a key driver in the evolution of mobile networks. This reduction not only applies to the pre-operational state (e.g. the deployment phase of a new network), but to the operational state as well. During the operational state, self-optimization processes can be performed to reduce the operating expenses (OPEX) of telecommunications operators. In this paper, we propose an agent-based mechanism for auto-tuning mobile networks with the aim of achieving energy savings in access networks. Firefly-based synchronization is used for the coordination of mutually dependent software agents located at the operator’s base stations. Once synchronized, agents can act together to obtain a global goal – i.e. the reduction of the telecommunications operator’s costs without reducing the quality of the provided services for mobile users. These agents build a self-organized overlay network where neither centralized nor decentralized control is needed and, therefore, is more robust.

Introduction

Reducing operating expenses (OPEX) is a key driver in the evolution of mobile networks [15], [22], [42]. Consequently, it is of vital interest to telecommunications operators to explore methods to minimize OPEX. One of the possible solutions is introducing self-optimizing mechanisms, which are aimed at increasing network performance and quality by autonomously reacting to dynamic processes in the network. It is important to highlight that a self-optimizing mobile network (SOMN) should auto-tune (i.e. autonomously tune) while in its operational state.

If auto-tuning of the mobile network is employed, one important challenge must be faced. Namely, distributed entities in the autonomous system must be successfully synchronized. In this paper, we address this challenge by proposing software agent technology as a tool for auto-tuning the network, while synchronization among distributed agents is achieved using swarm-intelligence [33], i.e. simulation of firefly behavior. Section 2 of this paper introduces the concept of software agents in the SOMN architecture. Section 3 elaborates upon synchronization in distributed systems, presenting different overlay network topologies of distributed agents. Section 4 presents a SOMN case study, where firefly-synchronized software agents auto-tune a mobile network with the aim of achieving energy savings in the access network. Finally, Section 5 concludes the paper and gives directions for future work.

Section snippets

Software agents for auto-tuning of the mobile network

Firefly Agents (FAs) are a vital part of the proposed system which automates actions on SOMN nodes [26], [28]. They are placed at each node in the SOMN and use firefly-based synchronization mechanisms to coordinate actions (see Fig. 1).

Generally, a software agent [4], [7], [14] is a program which autonomously acts on behalf of its organizational principal while carrying out information and communication tasks which have been delegated to it [27], [29]. From the owner’s (i.e. telecommunications

Synchronization in distributed systems

Understanding self-organized biological processes (e.g. trail formation in ants, bird flocking, fish schooling, pattern formation in bacteria, colony thermoregulation in honey bees) can be used as an inspiration for solving problems in various computer-supported domains (e.g. engineering or telecommunications). This kind of aggregate motion in biological systems is called “swarm behavior” and its application in computer-supported systems is referred as swarm intelligence (SI). Successful

Case study: enabling energy savings in mobile access networks

A stunning fact is that the total CO2 emission produced by the information and communication technology (ICT) industry exceeds that of the aviation industry. Given the fact that in less developed countries, the ICT industry has been growing rapidly, ICT is becoming a major consumer of energy in the world. It is expected that the growth of the total energy consumption, and thus the CO2 emission will continue to grow further in the near future. According to recent estimates, ICT consumes between

Conclusion and future work

In this paper a system based on software agent technology used to auto-tune the network is proposed, while synchronization of distributed agents is achieved using swarm-intelligence based on firefly behavior. The system consists of firefly agents (FAs) placed at each node in the network which are responsible for synchronization and coordination.

Different overlay network topologies are tested (mesh (k), ring, line and star) to find the most suitable one. Simulation results of the proposed

Acknowledgements

This work was carried out within research project 036-0362027-1639 “Content Delivery and Mobility of Users and Services in New Generation Networks”, supported by the Ministry of Science, Education and Sports of the Republic of Croatia.

Additionally, authors are very grateful to Tomislav Lipic and Kresimir Pripuzic, PhD for helping with the computer modeling and for many helpful comments and suggestions on their work. Authors are also very thankful to the anonymous reviewers for their many

References (48)

  • I. Bojic, M. Kusek, Fireflies synchronization in small overlay networks, in: Proceedings of 32nd International...
  • J.M. Bradshaw

    Software Agents

    (1997)
  • S. Camazine et al.

    Self-Organization in Biological Systems

    (2003)
  • D.N. Chorafas

    Agent Technology Handbook

    (1998)
  • W.T. Cockayne et al.

    Mobile Agents

    (1998)
  • J. Edmonds et al.

    Matching: a well-solved class of linear programs

    (1969)
  • Ericsson, Vodafone Germany first to launch Ericsson’s power-saving feature to reduce energy consumption and cut CO2...
  • Ericsson, Energy-saving solutions helping mobile operators meet commercial and sustainability goals worldwide, Press...
  • J.J. Hopfield et al.

    Rapid local synchronization of action potentials: toward computation with coupled integrate-and-fire neurons

    Proceedings of the National Academy of Sciences USA

    (1995)
  • N. Jennings et al.

    A roadmap of agent research and development

    Journal of Autonomous Agents and Multi-Agent Systems

    (1998)
  • K. Knightson et al.

    NGN architecture: generic principles, functional architecture, and implementation

    IEEE Communications Magazine

    (2005)
  • R. Leidenfrost, W. Elmenreich, Establishing wireless time-triggered communication using a firefly clock synchronization...
  • R. Leidenfrost, W. Elmenreich, Firefly clock synchronization in an 802.15.4 wireless network, EURASIP Journal on...
  • D. Lucarelli, I. Wang, Decentralized synchronization protocols with nearest neighbor communication, in: Proceedings of...
  • Cited by (41)

    • Potential bias when creating a differential-vector movement algorithm

      2021, Applied Soft Computing
      Citation Excerpt :

      A possible design approach for constructing differential-vector movement algorithms is shown in Fig. 1. Most proposals primarily focused on the last three activities in Fig. 1 to verify algorithmic properties of translational, rotational, and scale invariance [29–31], to balance search abilities of exploitation versus exploration [18–33], and to make comparisons over various optimization problems [34–36]. Few of these proposals considered the former five activities in the flowchart in Fig. 1.

    • A corrected and improved symbiotic organisms search algorithm for continuous optimization

      2021, Expert Systems with Applications
      Citation Excerpt :

      Each metaheuristic algorithm has specific characteristics that are suited to solving particular numerical optimization problems that are non-linear, non-differentiable, and complex. Examples of metaheuristic algorithms include genetic algorithm (GA) (Cheng et al., 2012), particle swarm optimization (PSO) (Eberhart and Kennedy, 1995; Kennedy and Eberhart, 1995; Tsai et al., 2012), fish swarm algorithm (Shen et al., 2011; Tsai and Lin, 2011), artificial bee colony (ABC) (Basturk and Karaboga, 2006; Karaboga et al., 2012), bees algorithm (Pham et al., 2005; Tsai, 2014a), gravitational search algorithm (Rashedi et al., 2009; Tsai et al., 2013), cuckoo search (Yang and Deb, 2009), firefly algorithm (Bojic et al., 2012), and symbiotic organism search (SOS) (Cheng and Prayogo, 2014), among others (Tsai, 2015; Balochian & Baloochian, 2019; Anita et al., 2020; Got et al., 2020; Gupta et al., 2020; Wu et al., 2020; Yimit et al., 2020). PSO has received strong support and attention in recent decades because its execution uses simple formulations and it converges rapidly on a solution.

    • Artificial bee colony directive for continuous optimization

      2020, Applied Soft Computing Journal
      Citation Excerpt :

      SI algorithms mostly simulate the natural behavior of biological life in settings such as colonies of ants, flocks of birds, herds of animals, bacterial growth, and schools of fish. SI algorithms include the particle swarm optimization (PSO) [3–5], fish swarm algorithm [6,7], artificial bee colony (ABC) [8–14], bees algorithm [15,16], gravitational search algorithm [17,18], cuckoo search [19], firefly algorithm [20], and so on [21]. PSO has received strong support and attention in recent years because of its execution using simple formulations and rapid converge on a solution.

    • Confined teaching-learning-based optimization with variable search strategies for continuous optimization

      2019, Information Sciences
      Citation Excerpt :

      SI-based algorithms simulate the natural behavior of living organisms such as colonies of ants, flocks of birds, herds of animals, bacterial growth, and schools of fish. Such algorithms include particle swarm optimization (PSO) [19], fish swarm algorithm [34], artificial bee colony (ABC) [2], bees algorithm [36], gravitational search algorithm [35], cuckoo search (CS) [42], and firefly algorithm [4]. PSO has received strong support and attention in recent decades because its execution uses simple formulations and because it converges rapidly on a solution.

    View all citing articles on Scopus
    View full text