When solving complex combinatorial problems such as the formation of traffic networks, standard mathematical algorithms often fall short. As a solution, algorithms inspired by nature and biology are becoming increasingly important. One of the best-studied organisms in this field is the slime mold Physarum polycephalum. It exhibits complex decision-making, associative learning, spatial and temporal memory, can solve mazes by finding the shortest path, and can be used to make logic gates, electronic components, and transportation networks. In this thesis, we have provided a detailed description of the biological basis for these properties and investigated the use of slime mold to improve the existing transportation network using Slovenian railroad as an example. We assumed that the slime mold is capable of forming an efficient network comparable to, or even better than, the Slovenian railroad network. The experiment confirmed our hypothesis, as the slime mold formed a network with the same travel speed as the Slovenian railroad but with more than twice the resilience to failures, albeit being 29% longer. However, a direct comparison is difficult because certain constraints affecting the formation of railroad networks are not taken into account. Therefore, transferring slime mold's behavior into in silico models has potential for developing effective algorithms to solve various problems in the fields of transportation, computer science, information technology, and engineering.
|
