Fig 1.Happiness Model
2-1 INSPIRATION
In a company, some personals attempt to obtain a new positon instead of
old position, which leads causes to be happiness. By changing, the
behavior of personal in the company we can modeled to an algorithm,
which can be one of the specific branch of swarm intelligence.
The model by four behaviors,which mentioned before, can be make decision
in the space of problem. The model should take into account the
balancing exploration and exploitation in the space of problems. Four
agent based on mentioned behaviorare indicated and introduced as follows
Ps H=the history of personal
Ps C= the current personal
Ps\({\ N}_{1}\)= the best neighbor of current personal
Ps \(G\) = the best personal in company or companies
The cost valuesof eachagentsare converted to the fitness value, which is
defined in general equation 2. After computing, we calculate probability
value for each agent, whichis given in the equation3.
\(\text{Fitness}_{i}\)=Exp
(\(\frac{f_{i}}{\text{Average}(F)})\)F=\({\{f}_{1}\)…\(f_{i}\ldots f_{n}\)}n=4(2)
\(P_{i}\)= \(\frac{\text{Fit}_{i}}{\sum_{i=1}^{n}\text{Fit}_{i}}\) (3)
The space of movement agents determined with twocriterionfor balancing
the exploration and exploitation and scape of local minimum.
First criterion:
If the cost value of the first neighbor is less than the cost value
ofthe current agentsatisfied it leads to two statesbased on the
generated uniformvalue including coefficient formula.
Rr1: uniformly distributed random number in the interval (0,1) (what
is happen in the world)
L: current iteration
n: number of agents
pn=n*0.75Sum=\(\sum_{i=1}^{\text{pn}}p_{i}\)(4)\(\mu_{1}\)=2-L*((2)/iteration number) (5)\(\mu_{2}\)=1-L*((1)/iteration number) (6)\(\varnothing\ \)=1-L*((1)/iteration number)
(7)alpha=0.6-l*((0.6-0.09)/iteration number) (8)