An Estimation of Fitness Evaluation in Evolutionary Algorithm for the Rectilinear Steiner Tree Problem

직각거리 스타이너 나무 문제의 하이브리드 진화 해법에서 효율적인 적합도 추정에 관한 연구

The rectilinear Steiner tree problem is to find a minimum-length rectilinear interconnection of a set of terminals in the plane. It is well known that the solution to this problem will be the minimal spanning tree (MST) on some set Steiner points. A hybrid evolutionary algorithm is introduced based upon the Prim algorithm. The Prim algorithm for the fitness evaluation requires heavy calculation time. The fitness value of parents is inherited to their child and the fitness value of child is estimated by the inherited structure of tree. We introduce four alternative evolutionary algorithms, Experiment result shows that the calculation time is reduced to 25% without loosing the solution quality by using the fitness estimation.