Metaheuristic Optimization in Seismic Structural Design and Inspection Scheduling of Buildings - Structural Seismic Design Optimization and Earthquake Engineering - page 168

Figure 5. City of Patras - Subdivision into struc-
tural blocks and the mean damage level distributed
over the structural blocks

Table 10. Damage Factor (DF) corresponding to
each damage level

Damage level Damage Factor (DF)

0 1.0

1 1.2

2 1.5

3 2.0

CONCLUSION

In this study the application of metaheuristic op-
timization and in particular Evolution Strategies,
Particle Swarm Optimization and Ant Colony
Optimization is examined in two problems of
great significance, the structural seismic design
optimization problem and the inspection schedul-
ing problem after a seismic hazard attack.
In the first problem examined in this study it
was found that with reference to the factors in-
fluencing the life-cycle cost estimation it can be
concluded that 10 to 20 records are not enough to
obtain reliable life-cycle cost analysis prediction
results. The structural type of the building affects
its structural performance. It has been verified that
a symmetrical structure sustains less damage and
therefore less repair cost during its life compared
to a non-symmetric structure. In both test examples
the effect of the other sources of uncertainty like
material properties, damping and mass proper-
ties is very significant varying considerably the
mean, the standard deviation and the fractiles of
the seismic response. Neglecting the influence of
modeling uncertainties (i.e. material properties and
design variables) in the prediction of the seismic
response can significantly underestimate the val-
ues of the seismic damage indices considered. As
a result the estimated value of the life cycle cost
varies considerably (up to 30%) compared to the
case where the cumulative impact of all sources
of randomness is considered. Furthermore, it has
been shown that designs obtained in accordance
to the European seismic design code are more

optimum allocation problem for the two different
numbers of inspection crews.
In the second step, the inspection prioritization
problem defined in Eq. (6) is solved by means of
the Ant Colony Optimization algorithm. Figures
8(a) and 8(b) depict the optimum routes achieved,
corresponding to the least time consuming route
required for each inspection group imitating from
their base. The base is the same for every inspec-
tion crew. The distances for the first and second
group are 17121 and 31540 respectively for the
two inspection groups while for the four are 9633.7,
10939, 11383 and 15740.
Figure 9 depicts the convergence histories of
the ACO algorithm. The vertical axis is the
minimum distance path among the ants for every
iteration.