The influence of optimization target selection on the structure of arterial tree models generated by constrained constructive optimization

Abstract

The computational method of constrained constructive optimization was used to generate complex arterial model trees by optimization with respect to a target function. Changing the target function also changes the tree structure obtained. For a parameterized family of target functions a series of trees was created, showing visually striking differences in structure that can also be quantified by appropriately chosen numerical indexes. Blood transport path length, pressure profile, and an index for relative segment orientation show clear dependencies on the optimization target, and the nature of changes can be explained on theoretical grounds. The main goal was to display, quantify, and explain the structural changes induced by different optimization target functions.