1: \begin{abstract}

2: We present two results on slime mold computations. In wet-lab experiments (Nature'00)

3: by Nakagaki et al. the slime mold Physarum polycephalum demonstrated its ability to solve shortest 

4: path problems. Biologists proposed a mathematical model, a system of differential equations, for 

5: the slime's adaption process (J. Theoretical Biology'07). It was shown that the process convergences 

6: to the shortest path (J. Theoretical Biology'12) for all graphs. We show that the dynamics actually

7: converges for a much wider class of problems, namely undirected linear programs with a non-negative 

8: cost vector. 

9: 

10: Combinatorial optimization researchers took the dynamics describing slime behavior as an inspiration

11: for an optimization method and showed that its discretization can $\epsilon$-approximately solve 

12: linear programs with positive cost vector (ITCS'16). Their analysis requires a feasible

13: starting point, a step size depending linearly on $\epsilon$, and a number of steps with quartic 

14: dependence on $\mathrm{opt}/(\epsilon\Phi)$, where $\Phi$ is the difference between the smallest cost

15: of a non-optimal basic feasible solution and the optimal cost ($\mathrm{opt}$). 

16: 

17: We give a refined analysis showing that the dynamics initialized with any strongly dominating 

18: point converges to the set of optimal solutions. Moreover, we strengthen the convergence

19: rate bounds and prove that the step size is independent of $\epsilon$, and the number of steps 

20: depends logarithmically on $1/\epsilon$ and quadratically on $\mathrm{opt}/\Phi$.

21: \end{abstract}

22: