Problem representation: In Genetic Algorithm, a pool/population of
strings is processed many times. Each member string in the pool/population
represents
a possible solution. Some strings represent infeasible solutions, while
some represent good solution. Eventually after many processing the
population converges to the best possible solution, i.e. only the copies
of good solutions are left and bad ones are eliminated.
In our case of 4 DOF freedom manipulator, a string would represent the
increments to the four joint angles; e.g. at a gives position, the best
string would give the best incremental values for the four joints.
"Best incremental values" refer to those increment values to the
joint angles which
would avoid obstacles and move towards the goal.
Thus a string(S) would look like: { d(theta1), d(theta2),
d(theta3), d(theta4) } .
Since here we directly coding the string a set of four real numbers, the process
would be called "Real-coded Genetic Algorithm".
If theta = { theta1, theta2,
theta3, theta4 }
represents the current state (joint angles) of the manipulator arm, then for this state,
it required to find a (S) which moves the manipulator towards the goal
by a small amount avoiding the obstacle. Thus reaching the goal involves a number of steps
and in each step we have to find the best (S).