Imperialist competitive algorithm .28Atashpaz-Gargari .26 Lucas 2007.29 List of metaphor-based metaheuristics

the imperialist competitive algorithm computational method used solve optimization problems of different types. of methods in area of evolutionary computation, ica not need gradient of function in optimization process. specific point of view, ica can thought of social counterpart of genetic algorithms (gas). ica mathematical model , computer simulation of human social evolution, while gas based on biological evolution of species.

this algorithm starts generating set of random candidate solutions in search space of optimization problem. generated random points called initial countries. countries in algorithm counterpart of chromosomes in gas , particles in particle swarm optimization (pso) , array of values of candidate solution of optimization problem. cost function of optimization problem determines power of each country. based on power, of best initial countries (the countries least cost function value), become imperialists , start taking control of other countries (called colonies) , form initial empires.

two main operators of algorithm assimilation , revolution. assimilation makes colonies of each empire closer imperialist state in space of socio-political characteristics (optimization search space). revolution brings sudden random changes in position of of countries in search space. during assimilation , revolution colony might reach better position , has chance take control of entire empire , replace current imperialist state of empire.

imperialistic competition part of algorithm. empires try win game , take possession of colonies of other empires. in each step of algorithm, based on power, empires have chance take control of 1 or more of colonies of weakest empire.

algorithm continues mentioned steps (assimilation, revolution, competition) until stop condition satisfied.

the above steps can summarized below pseudocode.

0) define objective function:



f
(

x

)
,


x

=
(

x

1


,

x

2


,
…
,

x

d


)
;



{\displaystyle f(\mathbf {x} ),\quad \mathbf {x} =(x_{1},x_{2},\dots ,x_{d});\,}


1) initialization of algorithm. generate random solution in search space , create initial empires.
2) assimilation: colonies move towards imperialist states in different in directions.
3) revolution: random changes occur in characteristics of countries.
4) position exchange between colony , imperialist. colony better position imperialist,
has chance take control of empire replacing existing imperialist.
5) imperialistic competition: imperialists compete take possession of colonies of each other.
6) eliminate powerless empires. weak empires lose power gradually , eliminated.
7) if stop condition satisfied, stop, if not go 2.

8) end