Citation: Zedadra, O.; Guerrieri, A.;
Seridi, H. LFA: A Lévy Walk and
Firefly-Based Search Algorithm:
Application to Multi-Target Search
and Multi-Robot Foraging. Big Data
Cogn. Comput. 2022, 6, 22. https://
doi.org/10.3390/bdcc6010022
Academic Editor: Moulay A.
Akhloufi
Received: 16 December 2021
Accepted: 17 February 2022
Published: 21 February 2022
Publisher’s Note: MDPI stays neutral
with regard to jurisdictional claims in
published maps and institutional affil-
iations.
Copyright: © 2022 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
big data and
cognitive computing
Article
LFA: A Lévy Walk and Firefly-Based Search Algorithm:
Application to Multi-Target Search and Multi-Robot Foraging
Ouarda Zedadra
1,
* , Antonio Guerrieri
2
and Hamid Seridi
1
1
LabSTIC, Department of Computer Science, 8 Mai 1945 University, P.O.Box 401, Guelma 24000, Algeria;
seridi.hamid@univ-guelma.dz
2
Institute for High Performance Computing and Networking (ICAR), CNR—National Research Council of
Italy, Via P. Bucci 8/9C, 87036 Rende, Italy; antonio.guerrieri@icar.cnr.it
* Correspondence: zedadra.ouarda@univ-guelma.dz
Abstract: In the literature, several exploration algorithms have been proposed so far. Among these,
Lévy walk is commonly used since it is proved to be more efficient than the simple random-walk
exploration. It is beneficial when targets are sparsely distributed in the search space. However,
due to its super-diffusive behavior, some tuning is needed to improve its performance, specifically
when targets are clustered. Firefly algorithm is a swarm intelligence-based algorithm useful for
intensive search, but its exploration rate is very limited. An efficient and reliable search could be
attained by combining the two algorithms since the first one allows exploration space, and the
second one encourages its exploitation. In this paper, we propose a swarm intelligence-based search
algorithm called Lévy walk and Firefly-based Algorithm (LFA), which is a hybridization of the
two aforementioned algorithms. The algorithm is applied to Multi-Target Search and Multi-Robot
Foraging. Numerical experiments to test the performances are conducted on the robotic simulator
ARGoS. A comparison with the original firefly algorithm proves the goodness of our contribution.
Keywords: swarm intelligence; swarm robotics; Lévy walk; Firefly algorithm; LFA; multi-target
search (MTS); Multi-Robot Foraging (MRF)
1. Introduction
In the future, Multi-Robots Systems (MRS) are intended to replace humans in some
tasks, such as exploration, search and rescue, monitoring, surveillance, cleaning, and so
on. To accomplish such tasks, mechanisms of distributed coordination and cooperation are
desirable for MRS. Swarm robotics is a new approach to the coordination of large numbers
of robots whose main inspiration stems from the observation of social insects [1]. Swarm
robotics has recently come out as the realizations of the swarm intelligence (SI) toward the
MRS. It emphasizes the incorporation of individuals and lifelike interactions, both among
individuals and between the individuals and their environment. Swarm robotics exposes a
synchronized behavior at the system level which emerges despite (i) the limited capabilities
of the individuals, (ii) the absence of centralized coordination, and (iii) the easiness of
interactions [2].
Firefly algorithm (FA) is an SI-based algorithm that takes advantage of lights to
coordinate fireflies, which, in nature, emit bioluminescent light to allure their mates or
preys. More is the brightness they emit, more is the attraction that they have. The light
intensity is proportional to the associated luminescence quantity called luciferin, which
attracts other fireflies within their neighborhood. The standard FA employs a full attraction
model, which results in oscillations during the search process, and low attraction can also
lead to premature convergence. Therefore, the number of attractions is very important [3].
The standard Firefly algorithm adopts a fixed step length which can cause it to become
trapped in the local optima, thus causing low precision. The exploration rate of FA is very
limited. FA is useful for MRS since its real application is guaranteed by Ambient Light
Big Data Cogn. Comput. 2022, 6, 22. https://doi.org/10.3390/bdcc6010022 https://www.mdpi.com/journal/bdcc