Abstract
We investigate solving capability of genetic algorithm (GA)
to an optimization problem of cable routing in a large scale building.
In solving the problem by GA,
in usual,
the problem is divided into two subproblems:
(i) one is an upper layer problem where one decides the order to
lay the cables corresponding to one of
combinatorial optimization problems,
and (ii) the other is an lower layer problem where
one lies the cables with the shortest route totally
corresponding to one of shortest path problems.
It is known well to solve the optimization problem of cable routing
by a multi step GA directly.
Therefore,
the purpose of this paper is to investigate
solving capability of GA to an optimization problem of cable routing
in realizing a practical multi step GA.
We apply GA to the upper layer problem of
the combinatorial optimization problem
with the change of the crossover method,
OBX and PBX, and the mutation method,
EM and IM.
From results,
the case of the crossover method, PBX,
and the mutation method, IM, gives the best result
from the viewpoint of the calculation speed.
Thus,
PBX and IM is practical
in solving the combinatorial optimization problem
of the order of the cables by GA.
