This paper describes the application of a genetic algorithm to the problem of minimising the weight of a double-layer space truss. The paper uses this example to address some of the important issues in the application of genetic algorithms including; identifying optimisation problems which are suited to a genetic algorithm, the need for the encoding system to support the schema theorem and building block hypothesis, identifying the critical minimum population size, choosing the optimal mutation and crossover rates.

The paper explains how highly constrained problems lead to long critical schema which may not be composed of shorter fit schemas and hence such problems may be difficult to solve using traditional genetic algorithms. The use of redundant heuristic operators to introduce schemas is also described along with the importance of operators which rehabilitate infeasible solutions.

In the example the genetic algorithm optimises the topology, shape and member sections of the double-layer grid subject to constraints on the overall deflection of the structure and the buckling of the compression members. Most double-layer grids are constructed from a three dimensional module which is repeated throughout the structure. The program developed optimises the basic module and the number of times it is repeated in each direction, i.e. north-south and east-west. The span in each direction and the number of modular repetitions define the module's plan dimensions. The coding also varies the depth of the module. The double layer grid is formed using three member sections for the top, three sections for the bottom and three sections for the intermediate members. The member sizes are selected using the GA while allocating an appropriate section to each member of the structure is based on iterative stress analyses.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £75 +P&P)