arXiv subjects
Publications and source records attributed to S. Redner.
This mini-review presents extensions of the voter model that incorporate various plausible features of real decision-making processes by individuals. Although these generalizations are not calibrated by empirical data, the resulting dynamics are suggestive of realistic collective social behaviors.
We determine how long a diffusing particle spends in a given spatial range before it dies at an absorbing boundary. In one dimension, for a particle that starts at $x_0$ and is absorbed at $x=0$, the average residence time in the range $[x,x+dx]$ is $T(x)=\frac{x}{D}\,dx$ for $x x_0$, where $D$ is the diffusion coefficient. We extend our approach to biased diffusion, to a particle confined to a finite interval, and to general spatial dimensions. We use the generating function technique to derive parallel results for the average residence time of the one-dimensional symmetric nearest-neighbor random walk that starts at $x_0=1$ and is absorbed at $x=0$. We also determine the distribution of times at which the random walk first revisits $x=1$ before being absorbed.
We investigate the motion of a run-and-tumble particle (RTP) in one dimension. We find the exact probability distribution of the particle with and without diffusion on the infinite line, as well as in a finite interval. In the infinite domain, this probability distribution approaches a Gaussian form in the long-time limit, as in the case of a regular Brownian particle. At intermediate times, this distribution exhibits unexpected multi-modal forms. In a finite domain, the probability distribution reaches a steady state form with peaks at the boundaries, in contrast to a Brownian particle. We also study the relaxation to the steady state analytically. Finally we compute the survival probability of the RTP in a semi-infinite domain. In the finite interval, we compute the exit probability and the associated exit times. We provide numerical verifications of our analytical results.
We study the dynamics of a \emph{myopic} forager that randomly wanders on a lattice in which each site contains one unit of food. Upon encountering a food-containing site, the forager eats all the food at this site with probability $p<1$; otherwise, the food is left undisturbed. When the forager eats, it can wander $\mathcal{S}$ additional steps without food before starving to death. When the forager does not eat, either by not detecting food on a full site or by encountering an empty site, the forager goes hungry and comes one time unit closer to starvation. As the forager wanders, a multiply connected spatial region where food has been consumed---a desert---is created. The forager lifetime depends non-monotonically on its degree of myopia $p$, and at the optimal myopia $p=p^*(\mathcal{S})$, the forager lives much longer than a normal forager that always eats when it encounters food. This optimal lifetime grows as $\mathcal{S}^2/\ln\mathcal{S}$ in one dimension and faster than a power law in $\mathcal{S}$ in two and higher dimensions.
We introduce the \emph{frugal foraging} model in which a forager performs a discrete-time random walk on a lattice, where each site initially contains $\mathcal{S}$ food units. The forager metabolizes one unit of food at each step and starves to death when it last ate $\mathcal{S}$ steps in the past. Whenever the forager decides to eat, it consumes all food at its current site and this site remains empty (no food replenishment). The crucial property of the forager is that it is \emph{frugal} and eats only when encountering food within at most $k$ steps of starvation. We compute the average lifetime analytically as a function of frugality threshold and show that there exists an optimal strategy, namely, a frugality threshold $k^*$ that maximizes the forager lifetime.
We investigate the absorption of diffusing molecules in a fluid-filled spherical beaker that contains many small reactive traps. The molecules are absorbed either by hitting a trap or by escaping via the beaker walls. In the physical situation where the number $N$ of traps is large and their radii $a$ are small compared to the beaker radius $R$, the fraction of molecules $E$ that escape to the beaker wall and the complementary fraction $T$ that eventually are absorbed by the traps depend only on the dimensionless parameter combination $λ= Na/R$. We compute $E$ and $T$ as a function of $λ$ for a spherical beaker and for beakers of other three-dimensional shapes. The asymptotic behavior is found to be universal: $1- E\sim λ$ for $λ\to 0$ and $E\simλ^{-1/2}$ for $λ\to\infty$.
This is a brief tutorial review of the dynamics of the voter model and the invasion process on complex networks.
We introduce a network growth model based on complete redirection: a new node randomly selects an existing target node, but attaches to a random neighbor of this target. For undirected networks, this simple growth rule generates unusual, highly modular networks. Individual network realizations typically contain multiple macrohubs---nodes whose degree scales linearly with the number of nodes $N$. The size of the network "nucleus"---the set of nodes of degree greater than one---grows sublinearly with $N$ and thus constitutes a vanishingly small fraction of the network. The network therefore consists almost entirely of leaves (nodes of degree one) as $N\to\infty$.
We investigate the dynamics of a greedy forager that moves by random walking in an environment where each site initially contains one unit of food. Upon encountering a food-containing site, the forager eats all the food there and can subsequently hop an additional $\mathcal{S}$ steps without food before starving to death. Upon encountering an empty site, the forager goes hungry and comes one time unit closer to starvation. We investigate the new feature of forager greed; if the forager has a choice between hopping to an empty site or to a food-containing site in its nearest neighborhood, it hops preferentially towards food. If the neighboring sites all contain food or are all empty, the forager hops equiprobably to one of these neighbors. Paradoxically, the lifetime of the forager can depend non-monotonically on greed, and the sense of the non-monotonicity is opposite in one and two dimensions. Even more unexpectedly, the forager lifetime in one dimension is substantially enhanced when the greed is negative; here the forager tends to avoid food in its local neighborhood. We also determine the average amount of food consumed at the instant when the forager starves. We present analytic, heuristic, and numerical results to elucidate these intriguing phenomena.
We investigate the role of greed on the lifetime of a random-walking forager on an initially resource-rich lattice. Whenever the forager lands on a food-containing site, all the food there is eaten and the forager can hop $\mathcal{S}$ more steps without food before starving. Upon reaching an empty site, the forager comes one time unit closer to starvation. The forager is also greedy---given a choice to move to an empty or to a food-containing site in its local neighborhood, the forager moves preferentially towards food. Surprisingly, the forager lifetime varies non-monotonically with greed, with different senses of the non-monotonicity in one and two dimensions. Also unexpectedly, the forager lifetime in one dimension has a huge peak for very negative greed.
How to best exploit patchy resources? This long-standing question belongs to the extensively studied class of explore/exploit problems that arise in a wide range of situations, from animal foraging, to robotic exploration, and to human decision processes. Despite its broad relevance, the issue of optimal exploitation has previously only been tackled through two paradigmatic limiting models---patch-use and random search---that do not account for the interplay between searcher motion within patches and resource depletion. Here, we bridge this gap by introducing a minimal patch exploitation model that incorporates this coupling: the searcher depletes the resources along its random-walk trajectory within a patch and travels to a new patch after it takes $\mathcal{S}$ consecutive steps without finding resources. We compute the distribution of the amount of resources $F_t$ consumed by time $t$ for this non-Markovian random walker and show that exploring multiple patches is beneficial. In one dimension, we analytically derive the optimal strategy to maximize $F_t$. We show that this strategy is robust with respect to the distribution of resources within patches and the criterion for leaving a given patch. We also show that $F_t$ can be optimized in the ecologically-relevant case of two-dimensional patchy environments.
We introduce a growing network model---the copying model---in which a new node attaches to a randomly selected target node and, in addition, independently to each of the neighbors of the target with copying probability $p$. When $p<\frac{1}{2}$, this algorithm generates sparse networks, in which the average node degree is finite. A power-law degree distribution also arises, with a non-universal exponent whose value is determined by a transcendental equation in $p$. In the sparse regime, the network is "normal", e.g., the relative fluctuations in the number of links are asymptotically negligible. For $p\geq \frac{1}{2}$, the emergent networks are dense (the average degree increases with the number of nodes $N$) and they exhibit intriguing structural behaviors. In particular, the $N$-dependence of the number of $m$-cliques (complete subgraphs of $m$ nodes) undergoes $m-1$ transitions from normal to progressively more anomalous behavior at a $m$-dependent critical values of $p$. Different realizations of the network, which start from the same initial state, exhibit macroscopic fluctuations in the thermodynamic limit---absence of self averaging. When linking to second neighbors of the target node can occur, the number of links asymptotically grows as $N^2$ as $N\to\infty$, so that the network is effectively complete as $N\to \infty$.
We introduce an evolving network model in which a new node attaches to a randomly selected target node and also to each of its neighbors with probability $p$. The resulting network is sparse for $p<\frac{1}{2}$ and dense (average degree increasing with number of nodes $N$) for $p\geq \frac{1}{2}$. In the dense regime, individual networks realizations built by this copying mechanism are disparate and not self-averaging. Further, there is an infinite sequence of structural anomalies at $p=\frac{2}{3}$, $\frac{3}{4}$, $\frac{4}{5}$, etc., where the dependences on $N$ of the number of triangles (3-cliques), 4-cliques, undergo phase transitions. When linking to second neighbors of the target can occur, the probability that the resulting graph is complete---where all nodes are connected---is non-zero as $N\to\infty$.
We investigate a stochastic search process in one, two, and three dimensions in which $N$ diffusing searchers that all start at $x_0$ seek a target at the origin. Each of the searchers is also reset to its starting point, either with rate $r$, or deterministically, with a reset time $T$. In one dimension and for a small number of searchers, the search time and the search cost are minimized at a non-zero optimal reset rate (or time), while for sufficiently large $N$, resetting always hinders the search. In general, a single searcher leads to the minimum search cost in one, two, and three dimensions. When the resetting is deterministic, several unexpected feature arise for $N$ searchers, including the search time being independent of $T$ for $1/T\to 0$ and the search cost being independent of $N$ over a suitable range of $N$. Moreover, deterministic resetting typically leads to a lower search cost than in stochastic resetting.
We study the dynamics of a starving random walk in general spatial dimension $d$. This model represents an idealized description for the fate of an unaware forager whose motion is not affected by the presence or absence of resources. The forager depletes its environment by consuming resources and dies if it wanders too long without finding food. In the exactly-solvable case of one dimension, we explicitly derive the average lifetime of the walk and the distribution for the number of distinct sites visited by the walk at the instant of starvation. We also give a heuristic derivation for the averages of these two quantities. We tackle the complex but ecologically-relevant case of two dimensions by an approximation in which the depleted zone is assumed to always be circular and which grows incrementally each time the walk reaches the edge of this zone. Within this framework, we derive a lower bound for the scaling of the average lifetime and number of distinct sites visited at starvation. We also determine the asymptotic distribution of the number of distinct sites visited at starvation. Finally, we solve the case of high spatial dimensions within a mean-field approach.
We determine the impact of resource renewal on the lifetime of a forager that depletes its environment and starves if it wanders too long without eating. In the framework of the minimal starving random walk model with resource renewal, there are three universal classes of behavior as a function of the renewal time. For sufficiently rapid renewal, foragers are immortal, while foragers have a finite lifetime otherwise. In one dimension, there is a third regime, for sufficiently slow renewal, in which the lifetime of the forager is independent of the renewal time. We outline an enumeration method to determine the mean lifetime of the forager in the mortal regime.
We investigate some simple and surprising properties of a one-dimensional Brownian trajectory with diffusion coefficient $D$ that starts at the origin and reaches $X$ either: (i) at time $T$ or (ii) for the first time at time $T$. We determine the most likely location of the first-passage trajectory from $(0,0)$ to $(X,T)$ and its distribution at any intermediate time $t<T$. A first-passage path typically starts out by being repelled from its final location when $X^2/DT\ll 1$. We also determine the distribution of times when the trajectory first crosses and last crosses an arbitrary intermediate position $x<X$. The distribution of first-crossing times may be unimodal or bimodal, depending on whether $X^2/DT\ll 1$ or $X^2/DT\gg 1$. The form of the first-crossing probability in the bimodal regime is qualitatively similar to, but more singular than, the well-known arcsine law.
We investigate the average time for the earliest particle to hit a spherical absorber when a homogeneous gas of freely diffusing particles with density $ρ$ and diffusivity $D$ is prepared in a deterministic state and is initially separated by a minimum distance $\ell$ from this absorber. In the high-density limit, this first absorption time scales as $\frac{\ell^2}{D}\frac{1}{\lnρ\ell}$ in one dimension; we also obtain the first absorption time in three dimensions. In one dimension, we determine the probability that the $k^{\rm th}$-closest particle is the first one to hit the absorber. At large $k$, this probability decays as $k^{1/3}\exp(-Ak^{2/3})$, with $A= 1.93299\ldots$ analytically calculable. As a corollary, the characteristic hitting time $T_k$ for the $k^{\rm th}$-closest particle scales as $k^{4/3}$; this corresponds to superdiffusive but still subballistic motion.