A. Lukina, A. Tiwari, S. Smolka, R. Grosu:

"Distributed adaptive-neighborhood control for stochastic reachability in multi-agent systems";

in: "SAC '19: Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing", Association for Computing Machinery, New YorkNYUnited States, 2019, ISBN: 978-1-4503-5933-7, ##.

We present

DAMPC

, a distributed, adaptive-horizon and adaptive-

neighborhood algorithm for solving the stochastic reachability prob-

lem in multi-agent systems, in particular, ocking modeled as a

Markov decision process. At each time step, every agent rst calls

a centralized, adaptive-horizon model-predictive control (AMPC)

algorithm to obtain an optimal solution for its local neighborhood.

Second, the agents derive the ock-wide optimal solution through

a sequence of consensus rounds. Third, the neighborhood is adap-

tively resized using a ock-wide cost-based Lyapunov function.

This way

DAMPC

improves eciency without compromising con-

vergence. We evaluate

DAMPC

īs performance using statistical model

checking. Our results demonstrate that, compared to AMPC,

DAMPC

achieves considerable speed-up (two-fold in some cases) with only

a slightly lower rate of convergence. The smaller average neighbor-

hood size and lookahead horizon demonstrate the benets of the

DAMPC

approach for stochastic reachability problems involving any

controllable multi-agent system that possesses a cost function.

Created from the Publication Database of the Vienna University of Technology.