The Big Bang-Big Crunch (BB-BC) algorithm is an optimization technique of swarm intelligence based on the Big Bang theory. It runs efficiently and can be realized simply

but it is easily trapped in local optimal results. For the purpose of overcoming its shortages

an improved BB-BC algorithm is put forward in this essay

with the changes of the reduced forms of blast radius

the distribution of the random variable

and the treatment of the boundary conditions. Besides

the improved algorithm is applied in damage detection of a simply supported beam with 10 and 20 elements respectively. The numerical simulations indicate that the identified results are excellent even in the great influence of noise

especially for successive elements with tiny damage. A conclusion can be drawn that the improved BBBC algorithm can precisely detect structure damage

and would not be easily trapped into local optimal.