Search

A key strategy for making production in factories more efficient is to collect data about the functioning of machines, and dynamically adapt their working. Such smart factories have data packets with a mix of stringent and non-stringent deadlines with varying levels of importance that need to be delivered via a wireless network. However, the scheduling of packets in the wireless network is crucial to satisfy the deadlines. In this work, we propose a technique of utilizing IEEE 802.11ax, popularly known as WiFi 6, for such applications. IEEE 802.11ax has a few unique characteristics, such as specific configurations of dividing the channels into resource units (RU) for packet transmission and synchronized parallel transmissions. We model the problem of scheduling packets by assigning profit to each packet and then maximizing the sum of profits. We first show that this problem is strongly NP-Hard, and then propose an approximation algorithm with a 12-approximate algorithm. Our approximation algorithm uses a variant of local search to associate the right RU configuration to each packet and identify the duration of each parallel transmission. Finally, we extensively simulate different scenarios to show that our algorithm works better than other benchmarks.

Flow sparsification is a classic graph compression technique which, given a capacitated graph $G$ on $k$ terminals, aims to construct another capacitated graph $H$, called a flow sparsifier, that preserves, either exactly or approximately, every multicommodity flow between terminals (ideally, with size as a small function of $k$). Cut sparsifiers are a restricted variant of flow sparsifiers which are only required to preserve maximum flows between bipartitions of the terminal set. It is known that exact cut sparsifiers require $2^{\Omega(k)}$ many vertices [Krauthgamer and Rika, SODA 2013], with the hard instances being quasi-bipartite graphs, where there are no edges between non-terminals. On the other hand, it has been shown recently that exact (or even $(1+\varepsilon)$-approximate) flow sparsifiers on networks with just 6 terminals require unbounded size [Krauthgamer and Mosenzon, SODA 2023, Chen and Tan, SODA 2024]. In this paper, we construct exact flow sparsifiers of size $3^{k^{3}}$ and exact cut sparsifiers of size $2^{k^2}$ for quasi-bipartite graphs. In particular, the flow sparsifiers are contraction-based, that is, they are obtained from the input graph by (vertex) contraction operations. Our main contribution is a new technique to construct sparsifiers that exploits connections to polyhedral geometry, and that can be generalized to graphs with a small separator that separates the graph into small components. We also give an improved reduction theorem for graphs of bounded treewidth [Andoni et al., SODA 2011], implying a flow sparsifier of size $O(k\cdot w)$ and quality $O\bigl(\frac{\log w}{\log \log w}\bigr)$, where $w$ is the treewidth.

In [{\em Generalized centers of finite sets in Banach spaces}, Acta Math. Univ. Comenian. (N.S.) {\bf 66}(1) (1997), 83--115], Vesel\'{y} developed the idea of generalized centers for finite sets in Banach spaces. In this work, we explore the concept of {\it restricted $\mr{F}$-center property} for a triplet $(X,Y,\mc{F}(X))$, where $Y$ is a subspace of a Banach space $X$ and $\mc{F}(X)$ is the family of finite subsets of $X$. In addition, we generalize the analysis to include all closed, bounded subsets of $X$. Similar to how Lindenstrauss characterized $n.2.I.P.$, we characterize $n.X.I.P.$. So, it is possible to figure out that $Y$ has $n.X.I.P.$ in $X$ for all natural numbers $n$ if and only if $\tr{rad}_Y(F)=\tr{rad}_X(F)$ for all finite subsets $F$ of $Y$. It then turns out that, for all continuous, monotone functions $f$, the $f$-radii viz. $\tr{rad}_Y^f(F),\tr{rad}_X^f(F)$ are same whenever the generalized radii viz. $\tr{rad}_Y(F), \tr{rad}_X(F)$ are also same, for all finite subsets $F$ of $Y$. We establish a variety of characterizations of central subspaces of Banach spaces. With reference to an appropriate subfamily of closed and bounded subsets, it appears that a number of function spaces and subspaces exhibit the restricted weighted Chebyshev center property.

Ranking algorithms find extensive usage in diverse areas such as web search, employment, college admission, voting, etc. The related rank aggregation problem deals with combining multiple rankings into a single aggregate ranking. However, algorithms for both these problems might be biased against some individuals or groups due to implicit prejudice or marginalization in the historical data. We study ranking and rank aggregation problems from a fairness or diversity perspective, where the candidates (to be ranked) may belong to different groups and each group should have a fair representation in the final ranking. We allow the designer to set the parameters that define fair representation. These parameters specify the allowed range of the number of candidates from a particular group in the top-$k$ positions of the ranking. Given any ranking, we provide a fast and exact algorithm for finding the closest fair ranking for the Kendall tau metric under block-fairness. We also provide an exact algorithm for finding the closest fair ranking for the Ulam metric under strict-fairness, when there are only $O(1)$ number of groups. Our algorithms are simple, fast, and might be extendable to other relevant metrics. We also give a novel meta-algorithm for the general rank aggregation problem under the fairness framework. Surprisingly, this meta-algorithm works for any generalized mean objective (including center and median problems) and any fairness criteria. As a byproduct, we obtain 3-approximation algorithms for both center and median problems, under both Kendall tau and Ulam metrics. Furthermore, using sophisticated techniques we obtain a $(3-\varepsilon)$-approximation algorithm, for a constant $\varepsilon>0$, for the Ulam metric under strong fairness., Comment: A preliminary version of this paper appeared in NeurIPS 2022

Ride-sharing is an essential aspect of modern urban mobility. In this paper, we consider a classical problem in ride-sharing - the Multi-Vehicle Dial-a-Ride Problem (Multi-Vehicle DaRP). Given a fleet of vehicles with a fixed capacity stationed at various locations and a set of ride requests specified by origins and destinations, the goal is to serve all requests such that no vehicle is assigned more passengers than its capacity at any point along its trip. We propose an algorithm HRA, which is the first non-trivial approximation algorithm for the Multi-Vehicle DaRP. The main technical contribution is to reduce the Multi-Vehicle DaRP to a certain capacitated partitioning problem, which we solve using a novel hierarchical grouping algorithm. Experimental results show that the vehicle routes produced by our algorithm not only exhibit less total travel distance compared to state-of-the-art baselines, but also enjoy a small in-transit latency, which crucially relates to riders' traveling times. This suggests that HRA enhances rider experience while being energy-efficient.

We continue to study (strong) property-$(R_1)$ in Banach spaces. As discussed by Pai \& Nowroji in [{\it On restricted centers of sets}, J. Approx. Theory, {\bf 66}(2), 170--189 (1991)], this study corresponds to a triplet $(X,V,\mathcal{F})$, where $X$ is a Banach space, $V$ is a closed convex set, and $\mathcal{F}$ is a subfamily of closed, bounded subsets of $X$. It is observed that if $X$ is a Lindenstrauss space then $(X,B_X,\mathcal{K}(X))$ has strong property-$(R_1)$, where $\mathcal{K}(X)$ represents the compact subsets of $X$. It is established that for any $F\in\mathcal{K}(X)$, $\textrm{Cent}_{B_X}(F)\neq\emptyset$. This extends the well-known fact that a compact subset of a Lindenstrauss space $X$ admits a nonempty Chebyshev center in $X$. We extend our observation that $\textrm{Cent}_{B_X}$ is Lipschitz continuous in $\mathcal{K}(X)$ if $X$ is a Lindenstrauss space. If $Y$ is a subspace of a Banach space $X$ and $\mathcal{F}$ represents the set of all finite subsets of $B_X$ then we observe that $B_Y$ exhibits the condition for simultaneously strongly proximinal (viz. property-$(P_1)$) in $X$ for $F\in\mathcal{F}$ if $(X, Y, \mathcal{F}(X))$ satisfies strong property-$(R_1)$, where $\mathcal{F}(X)$ represents the set of all finite subsets of $X$. It is demonstrated that if $P$ is a bi-contractive projection in $\ell_\infty$, then $(\ell_\infty, Range (P), \mathcal{K}(\ell_\infty))$ exhibits the strong property-$(R_1)$, where $\mathcal{K}(\ell_\infty)$ represents the set of all compact subsets of $\ell_\infty$. Furthermore, stability results for these properties are derived in continuous function spaces, which are then studied for various sums in Banach spaces.

Graph compression or sparsification is a basic information-theoretic and computational question. A major open problem in this research area is whether $(1+\epsilon)$-approximate cut-preserving vertex sparsifiers with size close to the number of terminals exist. As a step towards this goal, we study a thresholded version of the problem: for a given parameter $c$, find a smaller graph, which we call connectivity-$c$ mimicking network, which preserves connectivity among $k$ terminals exactly up to the value of $c$. We show that connectivity-$c$ mimicking networks with $O(kc^4)$ edges exist and can be found in time $m(c\log n)^{O(c)}$. We also give a separate algorithm that constructs such graphs with $k \cdot O(c)^{2c}$ edges in time $mc^{O(c)}\log^{O(1)}n$. These results lead to the first data structures for answering fully dynamic offline $c$-edge-connectivity queries for $c \ge 4$ in polylogarithmic time per query, as well as more efficient algorithms for survivable network design on bounded treewidth graphs., Comment: Merged version of arXiv:1910.10359 and arXiv:1910.10665 with improved bounds, 55 pages

In this paper, we initiate the study of fair clustering that ensures distributional similarity among similar individuals. In response to improving fairness in machine learning, recent papers have investigated fairness in clustering algorithms and have focused on the paradigm of statistical parity/group fairness. These efforts attempt to minimize bias against some protected groups in the population. However, to the best of our knowledge, the alternative viewpoint of individual fairness, introduced by Dwork et al. (ITCS 2012) in the context of classification, has not been considered for clustering so far. Similar to Dwork et al., we adopt the individual fairness notion which mandates that similar individuals should be treated similarly for clustering problems. We use the notion of $f$-divergence as a measure of statistical similarity that significantly generalizes the ones used by Dwork et al. We introduce a framework for assigning individuals, embedded in a metric space, to probability distributions over a bounded number of cluster centers. The objective is to ensure (a) low cost of clustering in expectation and (b) individuals that are close to each other in a given fairness space are mapped to statistically similar distributions. We provide an algorithm for clustering with $p$-norm objective ($k$-center, $k$-means are special cases) and individual fairness constraints with provable approximation guarantee. We extend this framework to include both group fairness and individual fairness inside the protected groups. Finally, we observe conditions under which individual fairness implies group fairness. We present extensive experimental evidence that justifies the effectiveness of our approach.

Given a graph $G=(V,E)$, capacities $w(e)$ on edges, and a subset of terminals $\mathcal{T} \subseteq V: |\mathcal{T}| = k$, a mimicking network for $(G,\mathcal{T})$ is a graph $(H,w')$ that contains copies of $\mathcal{T}$ and preserves the value of minimum cuts separating any subset $A, B \subseteq \mathcal{T}$ of terminals. Mimicking networks of size $2^{2^k}$ are known to exist and can be constructed algorithmically, while the best known lower bound is $2^{\Omega(k)}$; therefore, an exponential size is required if one aims at preserving cuts exactly. In this paper, we study mimicking networks that preserve connectivity of the graph exactly up to the value of $c$, where $c$ is a parameter. This notion of mimicking network is sufficient for some applications, as we will elaborate. We first show that a mimicking of size $3^c \cdot k$ exists, that is, we can preserve cuts with small capacity using a network of size linear in $k$. Next, we show an algorithm that finds such a mimicking network in time $2^{O(c^2)} \operatorname{poly}(m)$.

Given a graph $G=(V,E)$ with non-negative real edge lengths and an integer parameter $k$, the Min-Max k-Tree Cover problem seeks to find a set of at most $k$ subtrees of $G$, such that the union of the trees is the vertex set $V$. The objective is to minimize the maximum length among all the trees. We give the first constant factor approximation for the hard uniform capacitated version of this problem, where, an input parameter $\lambda$ upper bounds the number of vertices that can be covered by any of the trees. Our result extends to the rooted version of the problem, where we are given a set of $k$ root vertices, $R$ and each of the covering trees is required to include a distinct vertex in $R$ as the root. Prior to our work, the only result known was a $(2k-1)$-approximation algorithm for the special case when the total number of vertices in the graph is $k\lambda$ [Guttmann-Beck and Hassin, J. of Algorithms, 1997]. Our technique circumvents the difficulty of using the minimum spanning tree of the graph as a lower bound, which is standard for the uncapacitated version of the problem [Even et al., OR Letters 2004] [Khani et al., Algorithmica 2010]. Instead, we use Steiner trees that cover $\lambda$ vertices along with an iterative refinement procedure that ensures that the output trees have low cost and the vertices are well distributed among the trees.

In the Group Steiner Tree problem (GST), we are given a (vertex or edge)-weighted graph $G=(V,E)$ on $n$ vertices, a root vertex $r$ and a collection of groups $\{S_i\}_{i\in[h]}: S_i\subseteq V(G)$. The goal is to find a min-cost subgraph $H$ that connects the root to every group. We consider a fault-tolerant variant of GST, which we call Restricted (Rooted) Group SNDP. In this setting, each group $S_i$ has a demand $k_i\in[k],k\in\mathbb N$, and we wish to find a min-cost $H\subseteq G$ such that, for each group $S_i$, there is a vertex in $S_i$ connected to the root via $k_i$ (vertex or edge) disjoint paths. While GST admits $O(\log^2 n\log h)$ approximation, its high connectivity variants are Label-Cover hard, and for the vertex-weighted version, the hardness holds even when $k=2$. Previously, positive results were known only for the edge-weighted version when $k=2$ [Gupta et al., SODA 2010; Khandekar et al., Theor. Comput. Sci., 2012] and for a relaxed variant where the disjoint paths may end at different vertices in a group [Chalermsook et al., SODA 2015]. Our main result is an $O(\log n\log h)$ approximation for Restricted Group SNDP that runs in time $n^{f(k, w)}$, where $w$ is the treewidth of $G$. This nearly matches the lower bound when $k$ and $w$ are constant. The key to achieving this result is a non-trivial extension of the framework in [Chalermsook et al., SODA 2017], which embeds all feasible solutions to the problem into a dynamic program (DP) table. However, finding the optimal solution in the DP table remains intractable. We formulate a linear program relaxation for the DP and obtain an approximate solution via randomized rounding. This framework also allows us to systematically construct DP tables for high-connectivity problems. As a result, we present new exact algorithms for several variants of survivable network design problems in low-treewidth graphs., Comment: 28 pages, 2 figures

Online algorithms are usually analyzed using the notion of competitive ratio which compares the solution obtained by the algorithm to that obtained by an online adversary for the worst possible input sequence. Often this measure turns out to be too pessimistic, and one popular approach especially for scheduling problems has been that of "resource augmentation" which was first proposed by Kalyanasundaram and Pruhs. Although resource augmentation has been very successful in dealing with a variety of objective functions, there are problems for which even a (arbitrary) constant speedup cannot lead to a constant competitive algorithm. In this paper we propose a "rejection model" which requires no resource augmentation but which permits the online algorithm to not serve an epsilon-fraction of the requests. The problems considered in this paper are in the restricted assignment setting where each job can be assigned only to a subset of machines. For the load balancing problem where the objective is to minimize the maximum load on any machine, we give $O(\log^2 1/\eps)$-competitive algorithm which rejects at most an $\eps$-fraction of the jobs. For the problem of minimizing the maximum weighted flow-time, we give an $O(1/\eps^4)$-competitive algorithm which can reject at most an $\eps$-fraction of the jobs by weight. We also extend this result to a more general setting where the weights of a job for measuring its weighted flow-time and its contribution towards total allowed rejection weight are different. This is useful, for instance, when we consider the objective of minimizing the maximum stretch. We obtain an $O(1/\eps^6)$-competitive algorithm in this case. Our algorithms are immediate dispatch, though they may not be immediate reject. All these problems have very strong lower bounds in the speed augmentation model.

We describe a primal-dual framework for the design and analysis of online convex optimization algorithms for {\em drifting regret}. Existing literature shows (nearly) optimal drifting regret bounds only for the $\ell_2$ and the $\ell_1$-norms. Our work provides a connection between these algorithms and the Online Mirror Descent ($\omd$) updates; one key insight that results from our work is that in order for these algorithms to succeed, it suffices to have the gradient of the regularizer to be bounded (in an appropriate norm). For situations (like for the $\ell_1$ norm) where the vanilla regularizer does not have this property, we have to {\em shift} the regularizer to ensure this. Thus, this helps explain the various updates presented in \cite{bansal10, buchbinder12}. We also consider the online variant of the problem with 1-lookahead, and with movement costs in the $\ell_2$-norm. Our primal dual approach yields nearly optimal competitive ratios for this problem.