Testing Multiple Strategy Human Optimization based Artificial Human Optimization Algorithms
1
Satish Gajawada, 2Hassan M. H. Mustafa
1
Alumnus, Indian Institute of Technology Roorkee
1
Founder, Artificial Human Optimization – A New Field
2
Faculty of Specified Education, Dept. of Educational Technology, Banha University, Egypt
Abstract: Recently a new trend titled ‘Artificial Human Optimization’ has become popular in Evolutionary
Computing Domain. More than 30 papers were published in this new field proposed in December 2016.
‘Hassan Satish Particle Swarm Optimization (HSPSO)’ and ‘Human Inspired Differential Evolution (HIDE)’ are
the two latest Artificial Human Optimization algorithms proposed based on Multiple Strategy Human
Optimization. In this paper we focus on Testing HSPSO and HIDE by applying these latest algorithms on
Ackley, Bohachevsky, Booth, Three-Hump Camel and Beale benchmark functions. Results obtained for these
Artificial Human Optimization Algorithms are compared with Differential Evolution and Particle Swarm
Optimization.
Indexing terms/Keywords:Artificial Humans, Global Optimization Techniques, Artificial Human Optimization,
Nature Inspired Computing, Bio-Inspired Computing, Genetic Algorithms, Particle Swarm Optimization,
Differential Evolution, Evolutionary Computing
Date of Publication: 2018-08-30
Volume: 01 Issue: 02
Journal:Computer Reviews Journal
Website:https://purkh.com
This work is licensed under a Creative Commons Attribution 4.0 International License.
157
1. Introduction
In [1], Multiple Strategy Human Optimization (MSHO) was proposed. In MSHO, more than one strategy are
used for movement of Artificial Humans in search space. Artificial Humans move towards the best in even
generations and move away from the worst in odd generations. In [2], Human Inspired Differential Evolution
(HIDE) was proposed based on the concepts of MSHO and Differential Evolution. MSHO concepts are
incorporated into Particle Swarm Optimization to create Hassan Satish Particle Swarm Optimization (HSPSO)
[3]. But these algorithms are not completely tested. In this paper, HIDE and HSPSO are tested by applying
these latest algorithms on various benchmark functions. Section 2 shows HSPSO algorithm. HIDE algorithm is
shown in Section 3. Section 4 shows results obtained after applying HIDE and HSPSO on benchmark functions.
Section 5 shows the conclusion.
2. Hassan Satish Particle Swarm Optimization ( HSPSO )
Procedure: Hassan Satish Particle Swarm Optimization ( HSPSO )
1) Initialize all particles
2) iterations = 0
3) do
4) for each particle ido
5) If ( f( xi ) < f( pbesti) ) then
6) pbesti= xi
7) end if
8) if ( f( pbesti) < f( gbest ) ) then
9) gbest = pbesti
10) end if
11) If ( f( xi ) > f( pworsti) ) then
12) pworsti= xi
13) end if
14) if ( f( pworsti) > f( gworst ) ) then
15) gworst = pworsti
16) end if
17) end for
18) If ((iterations == 0) || (iterations%2==0)) then // for starting and even iterations
19) for each particle ido
158
20) for each dimension d do
21) vi,d = vi,d + C1*Random(0,1)*(pbesti,d – xi,d) + C2*Random(0,1)*(gbestd – xi,d)
22) xi,d= xi,d+ vi,d
23) end for
24) end for
25) else // for odd iterations
26) for each particle ido
27) for each dimension d do
28) vi,d = vi,d + C1*Random(0,1)*( xi,d - pworsti,d ) + C2*Random(0,1)*( xi,d - gworstd)
29) xi,d= xi,d+ vi,d
30) end for
31) end for
32) end if
33) iterations = iterations + 1
34) while( termination condition is false)
3. Human Inspired Differential Evolution ( HIDE )
Procedure : Human Inspired Differential Evolution ( HIDE )
1) begin
2) Set generation count G = 0.
Initialize population randomly for NP individuals.
Initialize parameters CR and F.
3) Calculate fitness for all individuals in the population.
4) while (termination condition not equals to true) do
5) for ( i = 1 to NP) do
6) Select 3 individuals Xa, Xb and Xc such that
7) Xi != Xa != Xb!= Xc
8) for ( j = 1 to D ) do
159
9) select jrand randomly from 1 to D
10) randno= rand(0,1)
11) if ( randno<= CR or j == jrand) then
12) If ( generation G is even || generation G is 0 ) then
/*
Moving towards the best individual in starting generation and even generations
*/
14) ui,j = Xbest,j + F * (Xb,j – Xc,j)
15) else
/*
Moving away from the worst individual in odd generations
*/
16) ui,j = Xa,j + F * (Xb,j – Xworst,j)
17)
18) end if
19) else
20) ui,j = Xi,j
21) end if
22) end for
23) end for
24) for ( i = 1 to NP ) do
25) calculate ui
26) if(ui is better than Xi) then
27) Xi = ui
28) end if
29) end for
30) Store the best solution achieved so far.
31) end while
160
32) end
4. Results
Results obtained after applying HSPSO and HIDE algorithms on various benchmark functions are shown in this
section. The figures and equations of benchmark functions are taken from [4].
Figure 1 Ackley Function
Figure 2 Equation of Ackley Function
Figure 3 Result given by HSPSO on Ackley Function
Figure 4 Result given by PSO on Ackley Function
161
From Figure 3 and Figure 4 it can be seen that PSO performed better than HSPSO on Ackley function.
Figure 5 Result given by HIDE on Ackley Function
Figure 6 Result given by DE on Ackley Function
From Figure 5 and Figure 6 it can be seen that HIDE and DE performed equally well on Ackley function.
Figure 7 Beale Function
Figure 8 Equation of Beale Function
Figure 9 Result given by HSPSO on Beale Function
162
Figure 10 Result given by PSO on Beale Function
From Figure 9 and Figure 10 it can be seen that PSO and HSPSO both performed well on Beale function.
Figure 11 Result given by HIDE on Beale Function
Figure 12 Result given by DE on Beale Function
From Figure 11 and Figure 12 it can be seen that HIDE and DE performed equally well on Beale function.
Figure 13 Bohachevsky Function
Figure 14 Equation of Bohachevsky Function
163
Figure 15 Result given by HSPSO on Bohachevsky Function
Figure 16 Result given by PSO on Bohachevsky Function
From Figure 15 and Figure 16 it can be seen that PSO performed better than HSPSO on Bohachevsky Function
Figure 17 Result given by HIDE on Bohachevsky Function
Figure 18 Result given by DE on Bohachevsky Function
From Figure 17 and Figure 18 it can be seen that HIDE and DE both performed equally well on Bohachevsky
Function.
Figure 19 Booth Function
164
Figure 20 Equation of Booth Function
Figure 21 Result given by HSPSO on Booth Function
Figure 22 Result given by PSO on Booth Function
From Figure 21 and Figure 22 it can be seen that HSPSO and PSO both performed well on Booth Function.
Figure 23 Result given by HIDE on Booth Function
Figure 24 Result given by DE on Booth Function
From Figure 23 and Figure 24 it can be seen that HIDE and DE both performed equally well on Booth Function.
Figure 25 Three-Hump Camel Function
165
Figure 26 Equation of Three-Hump Camel Function
Figure 27 Result given by HSPSO on Three-Hump Camel Function
Figure 28 Result given by PSO on Three-Hump Camel Function
From Figure 27 and Figure 28 it can be seen HSPSO and PSO both performed well on Three-Hump Camel
Function
Figure 29 Result given by HIDE on Three-Hump Camel Function
Figure 30 Result given by DE on Three-Hump Camel Function
From Figure 29 and Figure 30 it can be seen that HIDE and DE both performed equally well on Three-Hump
Camel Function
Conclusion
Human Inspired Differential Evolution (HIDE) and Differential Evolution (DE) performed equally on all
benchmark functions. Particle Swarm Optimization (PSO) performed better than Hassan Satish Particle Swarm
166
Optimization (HSPSO) on 2 benchmark functions. HSPSO and PSO both performed well on remaining
benchmark functions.
References
[1] Satish Gajawada, “Artificial Human Optimization – An Introduction”, Transactions on Machine Learning and
Artificial Intelligence, Volume 6, No 2, pp: 1-9, April 2018.
[2] Satish Gajawada, Hassan M. H. Mustafa , HIDE : Human Inspired Differential Evolution - An Algorithm under
Artificial Human Optimization Field , International Journal of Research Publications (Volume: 7, Issue: 1),
http://ijrp.org/paper_detail/264
[3] Satish Gajawada and Hassan M. H. Mustafa, Hybridizationconcepts of Artificial Human Optimization field
Algorithms incorporated into Particle Swarm Optimization (In Progress).
[4] https://www.sfu.ca/~ssurjano/ackley.html (accessed 28th July, 2018)
167
