A Method Inspired by Physarum for Solving the Steiner Problem
Atsushi Tero, Toshiyuki Nakagaki, Kazutaka Toyabe, Kenji Yumiki and Ryo Kobayashi
We propose a new solver for the Steiner tree problem, inspired by a true slime mold Physarum polycephalum. This problem involves finding the network that connects multiple points on a plane through the shortest total length. Such a network is known as the Steiner minimum tree (SMT). The solution of this problem is important for the design of transport and communication networks, but is not easy to obtain because the computational time required increases rapidly with the number of points. Using Melzak’s algorithm, it is almost impossible to find the best solution for more than thirty points. However, it is known that an amoeboid organism, Physarum plasmodium, can construct a network on an agar plate between many points at which food is placed. Because the Physarum network sometimes has the same topology as the SMT, we have studied how this is achieved by constructing a mathematical model for the network dynamics, based on the physiological mechanism. Our investigation enables us to propose and discuss the prospects of a new method for solving the Steiner problem.
Keywords: Physarum, slime mould, Steiner problem, adaptive network, combinatorial optimization problem, multi-purpose optimization.