Search

International audience; We consider the problem of periodic graph exploration in which a mobile entity with constant memory, an agent, has to visit all n nodes of an input simple, connected, undirected graph in a periodic manner. Graphs are assumed to be anonymous, that is, nodes are unlabeled. While visiting a node, the agent may distinguish between the edges incident to it; for each node v, the endpoints of the edges incident to v are uniquely identified by different integer labels called port numbers. We are interested in algorithms for assigning the port numbers together with traversal algorithms for agents using these port numbers to obtain short traversal periods. Periodic graph exploration is unsolvable if the port numbers are set arbitrarily; see Budach (1978) [1]. However, surprisingly small periods can be achieved by carefully assigning the port numbers. Dobrev et al. (2005) [4] described an algorithm for assigning port numbers and an oblivious agent (i.e., an agent with no memory) using it, such that the agent explores any graph with n nodes within the period 10n. When the agent has access to a constant number of memory bits, the optimal length of the period was proved in Gąsieniec et al. (2008) [7] to be no more than 3.75n−2 (using a different assignment of the port numbers and a different traversal algorithm). In this paper, we improve both these bounds. More precisely, we show how to achieve a period length of at most for oblivious agents and a period length of at most 3.5n−2 for agents with constant memory. To obtain our results, we introduce a new, fast graph decomposition technique called a three-layer partition that may also be useful for solving other graph problems in the future. Finally, we present the first non-trivial lower bound, 2.8n−2, on the period length for the oblivious case.

Burke, Carillo and Vakharia [2009] consider a class of single product sourcing problems with a stochastic demand and multiple uncertain suppliers. Assuming that the demand is independent of the supplier reliabilities and uniformly distributed, they propose to write the expected profit as a quadratic function and derive a closed form expression for the optimal orders. We show that this formula is true only under special circumstances, which are not satisfied in many practical situations of interest. We give an exact formulation and solution procedure, holding under general assumptions. We illustrate our point by a complementary analysis of the numerical examples given in the quoted paper.

International audience; Two anonymous mobile agents (robots) moving in an asynchronous manner have to meet in an infinite grid of dimension $\dime > 0$, starting from two arbitrary positions at distance at most $d$. Since the problem is clearly infeasible in such general setting, we assume that the grid is embedded in a $\dime$-dimensional Euclidean space and that each agent knows the Cartesian coordinates of its own initial position (but not the one of the other agent). We design an algorithm permitting the agents to meet after traversing a trajectory of length $O(d^\dime{\rm polylog\ }d)$. This bound for the case of {\sc 2d}~-grids subsumes the main result of \cite{CCGL}. The algorithm is almost optimal, since the $\Omega(d^\dime)$ lower bound is straightforward. Further, we apply our rendezvous method to the following network design problem. The ports of the $\dime$-dimensional grid have to be set such that two anonymous agents starting at distance at most $d$ from each other will always meet, moving in an asynchronous manner, after traversing a $O(d^\dime{\rm polylog\ }d)$ length trajectory. We can also apply our method to a version of the geometric rendezvous problem. Two anonymous agents move asynchronously in the $\dime$-dimensional Euclidean space. The agents have the radii of visibility of $r_1$ and $r_2$, respectively. Each agent knows only its own initial position and its own radius of visibility. The agents meet when one agent is visible to the other one. We propose an algorithm designing the trajectory of each agent, so that they always meet after traveling a total distance of $O((\frac{d}{r})^\dime {\rm polylog}(\frac{d}{r}))$, where $r = \min(r_1, r_2)$ and for $r\geq 1$.

We consider the task of exploring graphs with anonymous nodes by a team of non-cooperative robots, modeled as finite automata. For exploration to be completed, each edge of the graph has to be traversed by at least one robot. In this paper, the robots have no a priori knowledge of the topology of the graph, nor of its size, and we are interested in the amount of memory the robots need to accomplish exploration, We introduce the so-called {\em reduced automata technique}, and we show how to use this technique for deriving several space lower bounds for exploration. Informally speaking, the reduced automata technique consists in reducing a robot to a simpler form that preserves its “core” behavior on some graphs. Using this technique, we first show that any set of $q\geq 1$ non-cooperative robots, requires $\Omega(\log(\frac{n}{q}))$ memory bits to explore all $n$-node graphs. The proof implies that, for any set of $q K$-state robots, there exists a graph of size $O(qK)$ that no robot of this set can explore, which improves the $O(K^{O(q)})$ bound by Rollik (1980). Our main result is an application of this latter result, concerning {\em terminating} graph exploration with one robot, i.e., in which the robot is requested to stop after completing exploration. For this task, the robot is provided with a pebble, that it can use to mark nodes (without such a marker, even terminating exploration of cycles cannot be achieved). We prove that terminating exploration requires $\Omega(\log n)$ bits of memory for a robot achieving this task in all $n$-node graphs.