Biology

Slime molds efficiently solve complex routing problems like traveling salespeople

How the science connects

Optimization algor…Swarm intelligenceBiological computing

AI Insight

Researchers demonstrate that the slime mold Physarum polycephalum can solve traveling salesperson problems (TSPs) of up to eight cities by using optical feedback to guide its growth patterns. The organism's branches that represent correct solution paths exhibit distinctive characteristics: lower-frequency oscillations, higher amplitudes, and synchronized behavior compared to non-solution branches. These synchronization patterns emerge reliably and can predict the correct solution at the midpoint of the feedback interval with 100% accuracy across 41 trials, with the discrimination effects scaling predictably with problem size.


This work advances understanding of how biological systems can perform complex computational tasks without neurons or traditional computing hardware. The findings suggest practical methods for improving biocomputing approaches and reveal how organismal-scale coherence in living matter might be harnessed for information processing applications.


Understand the Science

Optimization algorithm Concept coming soon Swarm intelligence Concept coming soon Biological computing Concept coming soon

⚠️ Preprint – Noch nicht peer-reviewed

Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.

-cross
Abstract: The plasmodium of the true slime mold Physarum polycephalum-an ancient, unicellular, aneural organism-serves as a platform for studying the information-processing capacities of active matter. Previous experiments used Physarum’s intricate morphological dynamics and photoavoidance in stellate chips to solve $N$-city traveling salesperson problems (TSPs) of up to eight cities, scaling linearly in time with TSP size. Optical feedback controlled by a modified Hopfield network illuminated specific lanes at regular intervals, prompting Physarum to elongate or retract selected branches. When the illumination pattern stabilized in a non-equilibrium steady state, branches bifurcated reproducibly into solution and non-solution groups, with the former exhibiting lower-frequency, higher-amplitude, and more synchronized oscillations than the latter across 41 trials with valid TSP solutions. Physarum’s synchronization dynamics efficiently predict 100% of selected solutions by the midpoint of the optical-feedback interval, achieving statistically significant (paired t-test, $p<0.005$) discrimination from alternate tours well before the non-equilibrium steady state. Observed frequency downconversions and synchronized power amplifications scale linearly and quadratically, respectively, for small-to-moderate TSP size, as captured by a toy model of energy redistribution with saturating optical absorption. Tuning these features in native biomolecular chromophore networks may thus improve both the quality and efficiency of TSP solutions from Physarum-based biocomputers, which exploit the effects of organismal-scale coherence.

Source: Tracking and distinguishing slime mold solutions to traveling salesperson problems through synchronized amplification in the non-equilibrium steady state