A New Modified Firefly Algorithm

A New Modified Firefly Algorithm

PAPER A NEW MODIFIED FIREFLY ALGORITHM A New Modified Firefly Algorithm http://dx.doi.org/10.3991/ijes.v4i2.5879 Divya Gupta1, Medha Gupta2 1 Wize Commerce Private Limited, Gurgaon, India 2 Ambedkar Institute of Advanced Communication Technologies and Research (AIACTR), GGSIPU, New Delhi, India Abstract—Nature inspired meta-heuristic algorithms studies one optimum solution with a single solution space exists the emergent collective intelligence of groups of simple [8]. In these problems if a new solution has a better objec- agents. Firefly Algorithm is one of the new such swarm- tive function value than the old solution, the new solution based metaheuristic algorithm inspired by the flashing is accepted. An algorithm works in this space by accepting behavior of fireflies. The algorithm was first proposed in or rejecting the solutions based on their respective func- 2008 and since then has been successfully used for solving tion values. various optimization problems. In this work, we intend to Optimization problems based on swarm intelligence are propose a new modified version of Firefly algorithm known as meta-heuristic algorithms [9] which have fea- (MoFA) and later its performance is compared with the tures like self-organization, no central control, derivative standard firefly algorithm along with various other meta- free and easy to implement. These features lead to an heuristic algorithms. Numerical studies and results demon- emergent behavior that overcome the main limitations of strate that the proposed algorithm is superior to existing conventional methods and can be conveniently applied to algorithms. various optimization problems. These metaheuristic algo- Index Terms—Benchmark Functions, Firefly Algorithm, rithms have two major components namely exploitation Nature-Inspired Algorithms, Unconstrained Optimization and exploration. The exploration ability ensures that the algorithm can search the whole space and escape from I. INTRODUCTION local optima while the exploitation ability guarantees the algorithm can search carefully and converge to the opti- Swarm Intelligence is a collective behavior of decen- mal point [10]. tralized, self-organized natural or artificial systems. The Swarm Intelligence principles have been successfully concept was introduced by Gerardo Beni and Jing Wang applied in a variety of problem domains including func- in 1989 [1]. It is the study of computational systems in- tion optimization problems, finding optimal routes, sched- spired by the 'collective intelligence'. The collective intel- uling, structural optimization, and image and data analysis ligence emerges through the cooperation of large numbers [11] [12]. Computational modelling of swarms has been of homogeneous agents in the environment. The collective further applied to a wide-range of diverse domains, in- behavior of the organisms like ants [2], bees [3] [4], fish cluding machine learning [13], bioinformatics and medical [5] etc. inspires artificial intelligence to simulate and solve informatics [14], dynamical systems and operations re- real world problems. In such systems, there are a number search [15]; they have been even applied in finance and of simple agents which follow simple and fixed rules that business [16]. determine their possible behavior during interaction among themselves and the surroundings. Although the The firefly algorithm [17] is a swarm-based metaheuris- individual particles are ignorant of it, their collective be- tic algorithm which imitates the social behavior of fire- havior leads to an intelligent global behavior. In nature flies. Each firefly flashes its light with some brightness such systems are commonly used to solve problems such which attracts other fireflies within the neighborhood. The as effective foraging for food, prey evading, or colony re- closer the two fireflies are, the more attractive they will location. The information is typically stored throughout seem. When the attractiveness is proportional to the objec- the participating homogeneous agents, or is stored or tive function the search space is explored by moving the communicated in the environment itself such as through fireflies towards more attractive neighbors. The search the use of pheromones in ants, dancing in bees, and prox- strategy consists of controlled randomization, efficient imity in fish and birds. Like evolutionary computation, local search and selection of the best solutions. Though, swarm intelligence 'algorithms' or 'strategies' are consid- the algorithm has advantages of being easy to implement ered adaptive strategies and are typically applied to search and understand and has been successfully applied to vari- and optimization domains. ous engineering optimization problems since its emer- gence in 2008, it consists of disadvantages such as getting Optimization is the process of selecting the best ele- trapped in local optima or no memorizing capability. ment (with regard to some criteria) from some set of available alternatives or to find minimum or maximum The main aim of this paper is to eliminate all the exist- output for an experiment. The input consists of variables; ing limitations of Firefly algorithm by proposing a new the process or function is known as the cost function, the modified version of the same. Later a comprehensive objective function or the fitness function; and the output is comparative study on the performance of the proposed the cost or fitness [6]. An optimization problem [7] is a algorithm (MoFA) with the standard Firefly algorithm real world problem where the objective function is not (SFA) and other metaheuristic algorithms for optimizing a differentiable and its values can be acquired by simula- very large set of numerical functions is presented. tion. These problems are either single-objective or multi- objective. In single objective optimization problems only 18 http://www.i-jes.org PAPER A NEW MODIFIED FIREFLY ALGORITHM II. FIREFLY ALGORITHM gets associated with the objective function to be opti- mized. A. Standard Firefly Algorithm (SFA) In the Firefly algorithm, there are two important issues: Firefly algorithm was proposed by Dr. Xin She Yang in the variation of the light intensity and the formulation of 2008 and is inspired by the mating behavior of fireflies the attractiveness. As we know the intensity (I) of flashes [17]. Fireflies belong to the family of insects that are ca- decreases as the distance (r) increases, with the medium pable to produce natural light used to attract a mate or a given, it can be determined by equation . prey. There are about two thousand firefly species which ! Attractiveness (!) function is defined using an absorption!!! produce short and rhythmic flashes. These flashes often !!! ! ! !!! appear to be in a unique pattern and produce an amazing coefficient (!) and distance (r) by the equation sight in the tropical areas during summer. If a firefly is ! th th ! ! hungry or looks for a mate its light glows brighter in order The!!! Cartesian distance between any two i and j fire- to make the attraction of insects or mates more effective. ! flies! ! at !xi and xj respectively, the Cartesian distance is The brightness of the bioluminescent light depends on the available quantity of a pigment called luciferin, and more determined by equation where pigment means more light. Based on the bioluminescent th ! ! xi,k is the k component of!" the spatial!!! coordinate!!! !!! xi of the communication phenomenon of fireflies, the SFA uses th ! ! !! ! ! ! i firefly and d is the number of dimensions. three main basic rules: 1. A firefly will be attracted by other fireflies regardless of their sex. The movement of a firefly occurs when ith firefly gets attracted to another more attractive (brighter) jth firefly 2. Attractiveness is proportional to their brightness and which is determined by equation decreases as the distance among them increases. ! !!!!" 3. The landscape of the objective function determines . Here the second term is! due! to the! attraction! the brightness of a firefly [18]. while the third term is randomization!!! ! ! with! ! #! being! the! randomization!! ! !" parameter and being the vector of ran- Based on above three rules the pseudo code of SFA is dom numbers drawn from a Gaussian distribution. The prepared which is used by all the modified algorithms too. randomization parameter enhances! the searching ability of the algorithm whereas the attraction term ensures that the Pseudo code for SFA: attraction of a firefly to another is maximum when they Inputs: objective function f(x); variable boundary; are at the same place (i.e. r=0) and minimum when they are not in close proximity (i.e. r="). population size n; maximum attractiveness !0; absorp- tion coefficient "; randomization parameter # For any large number of fireflies (N), if N>>m, where m is the number of local optima of an optimization prob- Outputs: best solution lem, the convergence of the algorithm can be achieved. Here, the initial location of N fireflies is distributed uni- 1. Initialization; formly in the entire search space, and as the iterations of 2. For variable i = 1:n the algorithm continue fireflies converge into all the local 3. Randomly produce xi within the variable ranges; optimum. By comparing the best solutions among all these 4. End for i; optima, the global optima are achieved [18]. The parame- 5. Evaluate the function values of the firefly popu- ter ! characterizes the contrast of the attractiveness and its value varies from 0.1 to 10 determining the convergence

View Full Text

Details

  • File Type
    pdf
  • Upload Time
    -
  • Content Languages
    English
  • Upload User
    Anonymous/Not logged-in
  • File Pages
    6 Page
  • File Size
    -

Download

Channel Download Status
Express Download Enable

Copyright

We respect the copyrights and intellectual property rights of all users. All uploaded documents are either original works of the uploader or authorized works of the rightful owners.

  • Not to be reproduced or distributed without explicit permission.
  • Not used for commercial purposes outside of approved use cases.
  • Not used to infringe on the rights of the original creators.
  • If you believe any content infringes your copyright, please contact us immediately.

Support

For help with questions, suggestions, or problems, please contact us