Professional Documents
Culture Documents
Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
1. INTRODUCTION
Solar energy is a non-conventional type of energy. w
Solar energy has been harnessed by humans since where S is the solar constant (1369 ), r is the
ancient times using a variety of technologies. Solar m2
radiation, along with secondary solar-powered earth radius.
resources such as wave and wind power, The Handy formula which is used to calculate solar
hydroelectricity and biomass, account for most of energy received by earth
the available non-conventional type of energy on E 3.6 * (10 9 ) *S* n* r 2
earth. Only a small fraction of the available solar
energy is used.
Solar powered electrical generation relies on where E is the solar energy in EJ.
photovoltaic system and heat engines. Solar S is the Solar Constant in W/m2.
energy's uses are limited only by human creativity. n is the number of hours.
To harvest the solar energy, the most common way r is the Earth's radius in km
is to use photo voltaic panels which will receive 2. LITERATURE SURVEY
photon energy from sun and convert to electrical Due to protection failures during natural disaster,
energy. Solar technologies are broadly classified as the nuclear power plants are getting down. This
either passive solar or active solar depending on the lead to the concentration towards the renewable
way they detain, convert and distribute solar energy sources drastically. Among them, solar
energy. Active solar techniques include the use of power systems attract more attention because they
PV panels and solar thermal collectors to strap up provide excellent opportunity to generate electricity
the energy. while it is pollution and radiations free [1].
Passive solar techniques include orienting a However, solar energy systems generally suffer
building to the Sun, selecting materials with from the varying environmental condition and the
favorable thermal mass or light dispersing P–V characteristic curve exhibits a maximum
properties and design spaces that naturally circulate power point (MPP) that varies nonlinearly with
air Solar energy has a vast area of application such these conditions [2].
©www.ijrcct.org | 167
G. Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
©www.ijrcct.org | 168
G. Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
©www.ijrcct.org | 169
G. Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
©www.ijrcct.org | 170
G. Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
Although inertial mass, passive gravitational mass, Now, consider a system with N agents (masses).
and active gravitational mass are conceptually We define the position of the ith agent by:
distinct, no experiment has ever unambiguously
demonstrated any difference between them. The
X i xi1 ,..., xid ,..., xin for i=1,2,….,N
theory of general relativity rests on the assumption (4.6)
that inertial and passive gravitational mass are where xid represents the position of ith agent in the
equivalent. dth dimension.
This is known as the weak equivalence principle. At a specific time ‘t’, we define the force acting on
Standard general relativity also assumes the mass ‘i’ from mass ‘j’ as following:
equivalence of inertial mass and active
M pi t M aj t
gravitational mass; this equivalence is sometimes
called the strong equivalent principle. Fijd t G t
Rij t
x t x t
d
j
d
i
3.2 GRAVITATIONAL SEARCH ALGORITHM
(GSA) (12)
In this thesis, we introduce our optimization where Maj is the active gravitational mass related to
algorithm based on the law of gravity. In the agent j, Mpi is the passive gravitational mass related
proposed algorithm, agents are considered as to agent i, G(t) is gravitational constant at time t,
objects and their performance is measured by their is a small constant, and Rij(t) is the Euclidian
masses. All these objects attract each other by the distance between two agents i and j:
gravity force, and this force causes a global Rij t X i t , X j t
movement of all objects towards the objects with 2
heavier masses. Hence, masses cooperate using a (13)
direct form of communication, through To give a stochastic characteristic to our algorithm,
gravitational force. The heavy masses – which we suppose that the total force that acts on agent i
correspond to good solutions – move more slowly in a dimension d be a randomly weighted sum of
than lighter ones, this guarantees the exploitation dth components of the forces exerted from other
step of the algorithm. agents:
In GSA, each mass (agent) has four specifications:
position, inertial mass, active gravitational mass, N
and passive gravitational mass. The position of the Fi t
d
rand F t j
d
ij
mass corresponds to a solution of the problem, and j 1, j i
its gravitational and inertial masses are determined (14)
using a fitness function. In other words, each mass where randj is a random number in the interval [0,
presents a solution, and the algorithm is navigated 1].
by properly adjusting the gravitational and inertia Hence, by the law of motion, the acceleration of the
masses. By lapse of time, we expect that masses be agent i at time t, and in direction dth, aid(t) is given
attracted by the heaviest mass. This mass will as follows:
present an optimum solution in the search space.
Fi d t
The GSA could be considered as an isolated system aid t
of masses. It is like a small artificial world of M ii t
masses obeying the Newtonian laws of gravitation (15)
and motion. More precisely, masses obey the where Mii is the inertial mass of ith agent.
following laws: Furthermore, the next velocity of an agent is
Law of gravity: Each particle attracts every other considered as a fraction of its current velocity
particle and the gravitational force between two added to its acceleration. Therefore, its position and
particles is directly proportional to the product of its velocity could be calculated as follows:
their masses and inversely proportional to the
vid t 1 rand i vid t aid t
distance between them, R. We use here R instead
of R2, because according to our experiment results,
R provides better results than R2 in all experimental (16)
cases.
Law of motion: The current velocity of any mass is xid t 1 xid t vid t 1
equal to the sum of the fraction of its previous
velocity and the variation in the velocity. Variation (17)
in the velocity or acceleration of any mass is equal where randi is a uniform random variable in the
to the force acted on the system divided by mass of interval [0, 1]. We use this random number to give
inertia. a randomized characteristic to the search.
The gravitational constant, G, is initialized at the
beginning and will be reduced with time to control
©www.ijrcct.org | 171
G. Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
the search accuracy. In other words, G is a function initial value K0 at the beginning and decreasing
of the initial value ( G0 ) and time (t): with time. In such a way, at the beginning, all
Gt GG0 , t agents apply the force, and as time passes, K best is
(18) decreased linearly and at the end there will be just
Gravitational and inertia masses are simply one agent applying force to the others. Therefore,
calculated by the fitness evaluation. A heavier mass Eq. (19) could be modified as:
means a more efficient agent. This means that Fi d t rand j Fijd t
better agents have higher attractions and walk more jKbest , j i
slowly. Assuming the equality of the gravitational (26)
and inertia mass, the values of masses is calculated
using the map of fitness. We update the
where K best is the set of first K agents with the
gravitational and inertial masses by the following best fitness value and biggest mass.
equations: 3.3 ALGORITHM FOR GSA
M ai M pi M ii M i The steps of GSA are as follows:
Step 1: Initialization of Agents:
The positions of the N number of agents are
(19) initialized randomly.
fit i t worst t
X i xi1 ,..., xid ,..., xin for i=1,2,….,N
mi t xidrepresents the position of ith agent in the
best t worst t
where
th
d dimension.
(20) Step 2: Evolution of Fitness and computation:
mi t
M i t
For minimization or maximization problems, the
m j t
fitness evolution is performed by evaluating the
N
j 1 best and worst fitness for all agents at each
(21) iteration.
where fiti(t) represent the fitness value of the agent Minimization problems:
i at time t, and, worst(t) and best(t) are defined as bestt min fit j t
j1,..., N
follows (for a minimization problem):
bestt min fit j t
j1,..., N worstt max fit j t
j1,..., N
(22)
worstt max fit j t Maximization problems:
j1,..., N
(23) best t max fit j t
j1,..., N
best t max fit j t
j1,..., N
(24) worstt min fit j t
j1,..., N
worstt min fit j t where fiti(t) represent the fitness value of the agent
j1,..., N
(25) i at time t, and, worst(t) and best(t).
One way to perform a good compromise between Step 3: Calculation of G:
exploration and exploitation is to reduce the Gravitational constant G is computed at iteration t.
number of agents with lapse of time in Eq. (19). G t G0 e / T
Hence, we propose only a set of agents with bigger G0 and are initialized at the beginning and will
mass apply their force to the other. However, we be reduced with time to control
should be careful of using this policy because it the search accuracy. T is the total number of
may reduce the exploration power and increase the iterations.
exploitation capability. Step 4: Calculation of Masses of the agents:
We remind that in order to avoid trapping in a local Gravitational and inertia masses for each agent are
optimum the algorithm must use the exploration at calculated at iteration t.
beginning. By lapse of iterations, exploration must
fade out and exploitation must fade in. To improve
M ai M pi M ii M i
the performance of GSA by controlling exploration
and exploitation only the K best agents will attract
the others. K best is a function of time, with the
©www.ijrcct.org | 172
G. Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
fit i t worst t
mi t
best t worst t
mi t
M i t
m j t
N
j 1
Mai and Mpi are the active and passive gravitational
masses respectively, while Mii is the inertia mass of
the ith agent.
Step 5: Accelerations of agent’s calculation:
Acceleration of the ith agents at iteration t is
computed.
Fi d t
a t
d
M ii t
i
K best is the set of first K agents with the best Figure 4: Flow chart of GSA
fitness value and biggest mass. K best will decrease RESULTS
linearly with time and at the end there will be only
one agent applying force to the others. Fijd(t) is
computed as the following equation:
M pi t M aj t
Fijd t G t
Rij t
x t x t
d
j
d
i
©www.ijrcct.org | 173
G. Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
©www.ijrcct.org | 174
G. Lakshmi Hari Priyanka et al, International Journal of Research ISSN (Online) 2278- 5841
in computer and communication Technology , Vol 6, Issue- 5, May- 2017
ISSN (Print) 2320- 5156
©www.ijrcct.org | 175