call for papers, research paper publishing, where to publish research paper, journal publishing, how to publish research paper, Call For research paper, international journal, publishing a paper, IJEI, call for papers 2012,journal of science and technolog
This document summarizes a research paper that proposes using the firefly algorithm to solve the NP-hard optimization problem of assigning cells in a cellular network to switches in a way that minimizes total costs. The total cost includes handoff costs for calls between adjacent cells and cabling costs for connecting cells to switches, subject to constraints like switch capacity. It describes modeling the problem mathematically and comparing the firefly algorithm to the particle swarm optimization (PSO) algorithm. The firefly algorithm is shown to solve the cell-to-switch assignment problem faster than existing approaches like PSO.
Report
Share
Report
Share
1 of 6
Download to read offline
More Related Content
call for papers, research paper publishing, where to publish research paper, journal publishing, how to publish research paper, Call For research paper, international journal, publishing a paper, IJEI, call for papers 2012,journal of science and technolog
1. International Journal of Engineering Inventions
ISSN: 2278-7461, www.ijeijournal.com
Volume 1, Issue 2(September 2012) PP: 17-22
A Native Approach to Cell to Switch Assignment Using Firefly
Algorithm
Apoorva Sharma1, Shamsher Malik2
1
M.Tech Student, UIET, MDU, Rohtak, 2ECE Department, UIET, MDU, Rohtak
Abstract––This paper deals with a design problem for a network of Personal Communication Services (PCS). The goal is
to assign cells to switches in a PCS Network (PCSN) in an optimum manner so as to minimize the total cost which
includes two types of cost, namely handoff cost between two adjacent cells, and cable cost between cells and switches. The
problem of assigning cells to switches in a cellular mobile network is an NP-hard optimization problem. The design is to
be optimized subject to the constraint that the call volume of each switch must not exceed its call handling capacity.
However, because of the time complexity of the problem, the solution procedures are usually heuristic when the number
of cells and switches are more. In this paper we have proposed an assignment heuristic which is faster than the existing
algorithms. This paper describes Firefly Algorithm and proposes a possible way it can be adjusted to solve the problem of
assignment of cells of a geographical area to the available number of switches based on the minimization of total cost for
the assignment.
Keywords––PCS, cell to switch assignment, handoff, optimization, heuristics and combinatorial optimisation, PSO, FA.
I. INTRODUCTION
In 2008, Xin-She Yang [1] introduced a new Meta - heuristics algorithm Firefly Algorithm that is inspired by
firefly‟s social behaviour. In the basic form, this algorithm is designed to solve primarily continuous problems. The aim of
this work is to get an assignment of cells in a geographical area to the available switches so that the total cost of this
assignment is minimized. Since the last couple of decades, there have been significant advances in the development of
mobile communication systems. Even though significant improvement to communication infrastructure has been attained in
the personal communication service industry, the issues concerning the assignment of cells to switches in order to minimize
the cabling cost and handoff costs in a reasonable time still remain challenging [2].
The handoff caused by a subscriber movement from one cell to another, involves not only the modification of the
location of the subscriber in the database but also the execution of a fairly complicated protocol between switches S1 and S2
[3]. Therefore, there are two types of handoffs, one involves only one switch and the other involves two switches. The latter
is relatively more difficult and costly than the former. Intuitively, the cells among which the handoff frequency is high
should be assigned to the same switch as far as possible to reduce the cost of handoffs. However, since the call handling
capacity of each switch is limited, this should be taken as a constraint. Incorporating the cabling cost that occurs when a call
is connected between a cell and a switch, we have an optimization problem, called the cell assignment problem [4], of
assigning cells to switches such that the total cost, comprising of handoff cost between adjacent cells and cabling cost
between cells and switches, is minimized subject to the constraints of the call handling capacities of switches.
Here two algorithms are used to find this assignment:
1) PSO Algorithm
2) Firefly Algorithm
II. CELL ASSIGNMENT PROBLEM
In this section, the problem formulation and the mathematical modelling is discussed.
2.1 Problem Formulation
The assignment of cells to switches is first introduced by Arif Merchant and Bhaskar Sengupta [5] in 1995. This problem is
an NP-Complete problem. Based on these ideas presented above this paper presents the problem as:
“Assign all the cells in a particular geographical area to the available number of switches in order to minimize the total cost
which is the sum of cabling cost and handoff cost maintaining two constraints.”
The constraints for the problem are as follows:
1) Each cell must be assigned to exactly one switch.
2) Each switch has some limited capacity and assignment of cells must be done in such a way so that the total load on
the switch should not exceed the capacity of the switch.
17
2. A Native Approach To Cell To Switch Assignment Using Firefly Algorithm…
Fig 1: mapping representation for CSA problem [6]
III. MATHEMATICAL MODELLING
The assignment of cell to the switches is an NP-Hard problem, having an exponential complexity (n cells and m
switches) [2].
• Let no. of cells be „n‟ and no. of switches be „m‟
• hij – handoff cost between cell i and cell j
• cik – cabling cost between cell i and switch k
• dij – distance between cell i and switch (MSC) j
• Mk – call handling capacity of switch k
• λi - No of communication in cell i
• Yij – 1 if cell I and j are assigned to same switch and 0 otherwise.
• Xik – 1 if cell I is assigned to switch k and 0 otherwise.
For all cases, the range of i, j and k are defined as:
1 ≤ 𝑖 ≤ 𝑛, 1 ≤ 𝑗 ≤ 𝑛, 1 ≤ 𝑘 ≤ 𝑚
3.1 Formulation of constraints:
1) Each cell must be assigned to exactly one switch
𝑚
𝑥 𝑖𝑘 = 1, 1≤ 𝑖≤ 𝑚 … (1)
𝑘=1
2) Each switch has some capacity
𝑛
𝜆 𝑖 𝑥 𝑖𝑘 ≤ 𝑀 𝑘 , 1 ≤ 𝑘 ≤ 𝑚 … (2)
𝑖=1
3.2 Formulation of Cost Function:
1) Total Cabling Cost: this is formulated as a function of distance between base station and switch and number of calls that a
cell can handle per unit time [7]. 𝑐 𝑖𝑗 (𝜆 𝑗 ) is the cost of cabling per kilometre which is also modelled as a function of the
number of calls that a cell i can handle as:
𝑐 𝑖𝑗 = 𝐴 𝑖𝑗 + 𝐵 𝑖𝑗 𝜆 𝑗 … (3)
𝑚
𝑐 𝑖𝑗 𝜆 𝑗 𝑑 𝑖𝑗 𝑥 𝑖𝑗
𝑗 =1
𝑓𝑜𝑟 𝑖 = 1,2, … 𝑛 … (4)
2) Total handoff cost: we consider two types of handoffs, one which involves only one switch and another which involves
two switches. The handoff that occurs between cells that belong to the same switch consume much less network resources
than what occurs between cells that belongs to two different switches.
𝑛 𝑛
𝑖𝑗 (1 − 𝑦 𝑖𝑗 ) … (5)
𝑖=1 𝑗 =1
3) Total Cost: So our objective is to minimize the total cost which can be formed by the summation of all three costs. The
objective function is given by:
18
3. A Native Approach To Cell To Switch Assignment Using Firefly Algorithm…
𝑚 𝑛 𝑛
𝑐 𝑖𝑗 (𝜆 𝑗 )𝑑 𝑖𝑗 𝑥 𝑖𝑗 + 𝑖𝑗 (1 − 𝑦 𝑖𝑗 ) … (6)
𝑗 =1 𝑖=1 𝑗 =1
IV. EXISTING METHODOLOGY
There are various methods available for assigning cell to switches but this problem gets complicated for large size
hence heuristics are needed to solve this problem. In this paper Firefly algorithm is proposed and the results are compared
with PSO.
4.1 Particle Swarm Optimisation
PSO method consists of a collection (called swarm) of individual entities (called particles). Each particle
represents a solution in the solution space and iteratively evaluates the fitness of the candidate solutions and remembers the
location where they had their best success. The individual's best solution is called the local best (lbest). Each particle makes
this information available to other particles and could learn the global best solution (gbest). [8].
- A global best (gbest) is known to all and immediately updated when a new best position is found by any particle in the
swarm.
- The local best (lbest), which is the best solution that the particle has seen.
Based on these gbest and lbest values the velocity and the position of each particle is updated. The particle position and
velocity update equations in the simplest form that govern the PSO are given by [8].
𝑉𝑡+1 = 𝑐1 𝑉𝑡 + 𝑐2 𝑟1 𝑙𝑏𝑒𝑠𝑡 − 𝑥 𝑡 + 𝑐3 𝑟2 𝑔𝑏𝑒𝑠𝑡 − 𝑥 𝑡 … (7)
𝑥 𝑡+1 = 𝑥 𝑡 + 𝑉𝑡+1 … (8)
where, c1, c2, c3 are the constant weight vector for the previous velocity, local knowledge and global knowledge
respectively and r1 and r2 are two random variables in the range of 0 and 1 [2]. The selection of coefficients in the velocity
update equations affects the convergence and the ability of the swarm to find the optimum.
4.2 PSO Algorithm [2]
Step 1:
Initialize the objective function which is to be minimized and various user defined constants.
Initialize the number of particles.
Set the output or position of maximum or minimum in the objective function.
Step 2: Initialize all the particles and there corresponding positions or solution in the solution space.
Step 3: Find the best particles which has the minimum solution (for minima problems) or maximum solution (for maxima
problems). The solution of this particle is called global best (gbest).
Step 4: For each element find the local best solution which is the minimum solution up to that iteration. This local best
solution for each particle is called lbest.
Step 5: Now using lbest and gbest update the velocity of each particle using equation (7) and then update the position of each
particle using equation (8).
Step 6: Repeat steps 3 to 5 until stopping criterion met.
V. FIREFLY ALGORITHM
In this section the actual implementation of firefly algorithm for cell to switch assignment problem is explained
[9]. As discussed above that each firefly in the solution space is a complete solution which is independent to the other
fireflies. Implementation of firefly algorithm starts with spreading a particular number of fireflies in the solution space. The
initial position of fireflies may be random or fixed as given by the user. The initial solution means the initial assignment
matrix for each firefly and correspondingly the cost for that assignment [10]. Now each firefly in the solution space has an
assignment and corresponding cost for that assignment. Based on this cost the firefly with minimum cost is selected as the
brightest firefly and all the other fireflies move towards this brightest firefly depending upon randomness of firefly, distance
between fireflies and absorption coefficient of fireflies. This is called motion update for firefly. In the cell assignment
problem motion update means update the elements of assignment matrix. Thus the assignment matrix of each firefly is
updated according to the cost of each firefly.
Now for this new assignment again total cost is calculated for each firefly and all the parameters like brightest firefly and
update value for each assigned matrix are updated and this process is repeated until a stopping criterion is met [11]. From all
the cost values of each firefly the one which is minimum of all is selected as the best value and the corresponding
assignment is selected as the best assignment for the cells.
5.1 Implementation Details
The steps for implementation of firefly are as follows:
Step 1
Initialize the number of cells (n), switches (m) and number of fireflies (p) in the solution space.
Initialize position of cells and switches randomly in the search space.
Calculate distance between each cell and switch.
19
4. A Native Approach To Cell To Switch Assignment Using Firefly Algorithm…
Step 2
Generate the assigned matrix (xij) for each firefly where each particle is between 0 and 1.
The row of the matrix represents switches and column represents cells.
Step 3
Obtain solution matrix from the assigned matrix by making the largest value of each column to 1 and all other are
set to 0.
Step 4
Calculate the total cost based on this solution matrix.
Step 5
On the basis of this new cost the brightest firefly is found which has the minimum cost for the assignment.
Now update the position of all other fireflies based on the attractiveness of best firefly and also on the basis of
distance and randomness of fireflies.
Now update the position of best firefly randomly.
Repeat step 3 to 5 until stopping criterion is met.
VI. EXPERIMENTAL RESULTS AND ANALYSIS
All the experiments are done in MATLAB for various cases of cells and switches for two algorithms namely
firefly algorithm (FA) and Particle Swarm Optimization (PSO). The various parameters used for firefly algorithm are as
follows:
Randomness, α=1, absorption coefficient, γ=1, brightness at source=1.
The parameters used in the initialization of problem are: Handoff cost between two cells= 0 to 5,000 per hour,
constant A and B used in cabling cost=1 and0.001 respectively, call handling capacity of a switch=98000, number of
communication in a cell=500 to 1000 per hour.
The comparison of total cost and CPU time is shown below in graphs.
Table 1 Comparison Table
Switches, Cabling Handoff Total cost by Total cost by CPU Time CPU Time
cells cost cost FA PSO FA PSO
2, 25 116 1919 2036 2177 0.175 0.2655
2, 50 316 7723 8040 9070 0.1996 0.2843
2, 100 496 32840 33337 35339 0.2766 0.3472
2, 150 851 75953 76804 79461 0.3929 0.4456
2, 200 1277 135630 136910 140670 0.5763 0.67
2, 250 1440 212390 213530 2201330 0.7713 0.9318
3, 25 153 2317 2470 2989 0.1723 0.2644
3, 50 361 10728 11095 11436 0.1701 0.2579
3, 100 591 43593 44185 46553 0.2635 0.3231
3, 150 693 98487 99180 105210 1.0216 1.185
3, 200 977 181380 182360 187100 1.786 1.7541
3, 250 1536 281720 283260 291850 2.2914 2.5471
5, 25 144 2945 3089 3710 0.2138 0.2783
5, 50 337 12673 13010 13936 0.3079 0.3817
5, 100 507 54483 54990 56452 0.5929 0.633
5, 150 990 122020 123010 126100 1.1268 1.158
5, 200 964 219390 220360 224530 1.5611 1.5302
5, 250 1460 347340 348800 352100 2.3752 2.5678
10, 25 135 3493 3623 3980 0.2244 0.286
10, 50 346 14881 15227 15538 0.3364 0.38
10, 100 515 61153 61668 62940 0.66 0.7132
10, 150 837 140070 140910 142990 1.1328 1.1871
10, 200 1163 246800 247970 253440 1.9344 1.9383
10, 250 1535 391560 393100 394120 2.5006 2.5572
VII. CONCLUSION AND FUTURE WORK
The Firefly Algorithm is successfully implemented. From the figures (a-d) and (e-h) we can see that the cost in
two algorithms is comparable but the CPU time requirement in case of firefly is less as in case of PSO. As the iterations are
increased the probability of finding the lowest cost increases. But this will also increase the CPU time spent in finding this
minimum cost. The CPU time required for finding the minimum cost is comparatively less in Firefly algorithm as compared
20
5. A Native Approach To Cell To Switch Assignment Using Firefly Algorithm…
to PSO. Thus Firefly gives better performance when the execution time is concerned. In future, this work can be extended to
include switching and paging cost to the total cost for cell to switch assignment.
Cost Comparison Graphs:
90
80
70
60
East
50
40 West
30
North
20
10
0
1st Qtr 2nd Qtr 3rd Qtr 4th Qtr
The above graphs shows the cost comparison for the proposed Firefly Algorithm and PSO.
CPU Time Comparison Graphs:
21
6. A Native Approach To Cell To Switch Assignment Using Firefly Algorithm…
The above graphs show the CPU Time comparison graphs for proposed Firefly Algorithm and PSO.
REFRENCES
1. Xin-She Yang, “Nature-Inspired Metaheuristic Algorithms”, Luniver Press, 2008.
2. Siba K. Udgata, U. Anuradha, G. Pawan Kumar, Gauri K. Udgata, “Assignment of Cells to Switches in a Cellular
Mobile Environment using Swarm Intelligence”, IEEE International Conference on Information Technology, pp
189-194, 2008.
3. A. Merchant, B. Sengupta, “Assignment of cells to switches in PCS networks”, IEEE/ACM Transactions on
Networking, Vol. 3 (5) (1995), pp.521–526.
4. P. Bhattacharjee, D. Saha, A. Mukherjee, “Heuristics for Assignment of Cells to Switches in a PCSN: A
Comparative Study”, Intl. Conf. on Personal Wireless Communications, Jaipur, India, 1999, pp. 331–334.
5. Partha Sarathi Bhattacharjee, Debashis Saha, Amitava Mukherjee, “CALB: A New Cell to Switch Assignment
Algorithm with Load Balancing in the Design of a Personal Communication Services Network (PCSN), IEEE, pp
264-268, 2000.
6. Cassilda Maria Ribeiro Aníbal Tavares Azevedo Rodolfo Florence Teixeira Jr.,” Problem of Assignment Cells to
Switches in a Cellular Mobile Network” via Beam Search Method, WSEAS TRANSACTIONS on
COMMUNICATIONS,2010.
7. Shxyong JianShyua, B.M.T. Linb, Tsung ShenHsiaoa, “Ant colony optimization for the cell assignment problem
in PCS networks”, march, 2005.
8. James Kennedy, Russell Eberhart “Particle Swarm Optimization”, Proc. IEEE Int'l. Conf. on Neural
Networks(Perth, Australia), IEEE Service Center, Piscataway, NJ, pp.1942-1948, 1995.
9. Xin-She Yang, “Firefly Algorithm For Multimodal Optimization”, Luniver Press, 2008.
10. Xin-She Yang, “Metaheuristic optimization”, December 2010
11. K. Rajalakshmi, Prakash Kumar, Hima M Bindu “Targeting Cells to Switch Assignment of Cellular Mobile
Network using Heuristic Algorithm” LATEST TRENDS on COMPUTERS (Volume I),2011.
22