Title:Efficient Approximation Algorithms for the Largest Weight Data Retrieval Problem

Abstract:In mobile networks, wireless data broadcast is a powerful approach for disseminating a large number of data copies to a great number of clients. The largest weight data retrieval (LWDR) problem, first addressed by \cite{Infocom12LuEfficient}, is to schedule the download of a specified subset of data items in a given time interval, such that the weight of the downloaded items will be maximized.
In this paper, we present an approximation algorithm with ratio $0.632$ for LWDR via approximating the maximum sectionalized coverage (MSC) problem, a generalization of the maximum coverage problem which is one of the most famous ${\cal NP}-$complete problems. Let $\mathbb{F}=\{\mathbb{F}_{1},\dots,\mathbb{F}_{N}\}$, in which $\mathbb{F}_{i}=\{S_{i,\, j}\vert j=1,\,2\,\dots\}$ is a collection of subsets of $S=\{u_{1},\, u_{2},\,\dots,u_{n}\}$, and $w(u_{i})$ be the weight of $u_{i}$. MSC is to select at most one $S_{i,\, j_{i}}$ from each $\mathbb{F}_{i}$, such that $\sum_{u_{i}\in S'}w(u_{i})$ is maximized, where $S'=\underset{i=1}{\overset{N}{\cup}}S_{i,\, j_{i}}$. First, the paper presents a factor-$0.632$ approximation algorithm for MSC by giving a novel linear programming (LP) formula and employing the randomized LP rounding technique. By reducing from the maximum 3 dimensional matching problem, the paper then shows that MSC is ${\cal NP}-$complete even when every $S\in\mathbb{F}_{i}$ is with cardinality 2, i.e. $|S|=2$. Last but not the least, the paper gives a method transforming any instance of LWDR to an instance of MSC, and shows that an approximation for MSC can be applied to LWDR almost preserving the approximation ratio. That is a factor-$0.632$ approximation for LWDR, improving the currently best ratio of $0.5$ in \cite{Infocom12LuEfficient}.