This article addresses the top-k query processing problem on mobile ad hoc networks (MANETs). Top-k query processing is common to retrieve only highly important data items. However, methods for top-k query processing are not enough efficient and accurate in MANET environments. For improving the efficiency and accuracy, replication is a promising technique that each node in MANETs replicates data items retained by other nodes into its storage. Therefore, we fully combine the top-k query processing with data replication. We propose a framework that efficiently processes top-k queries based on a new replication strategy. We develop new replication strategy FReT (topology-Free Replication for Top-k query). FReT determines near-optimal allocations of replicas. It advantages efficient top-k query processing from limited search area without maintenance costs even if mobile nodes move. Our top-k query processing methods retrieve the exact answer with small overhead and delay by gradually increasing the search area based on FReT. We demonstrate, through extensive experiments, that FReT and query processing methods function well in terms of small delay and overhead without sacrificing exactness of the query result.

中文翻译：

---

在移动自组织网络中通过复制处理前k个查询的有效框架

本文解决了移动自组织网络（MANET）上的前k个查询处理问题。通常仅使用top-k查询处理来检索非常重要的数据项。但是，在MANET环境中，用于top-k查询处理的方法不够高效且不够准确。为了提高效率和准确性，复制是一种有前途的技术，MANET中的每个节点都将其他节点保留的数据项复制到其存储中。因此，我们将前k个查询处理与数据复制完全结合在一起。我们提出了一个基于新的复制策略有效处理前k个查询的框架。我们开发新的复制策略FReT（Top-k查询的无拓扑复制）。FReT确定副本的最佳分配。即使移动节点移动，它也可以从有限的搜索区域进行高效的前k查询处理，而无需维护成本。我们的前k个查询处理方法通过逐渐增加基于FReT的搜索区域，以较小的开销和延迟来检索准确的答案。通过大量的实验，我们证明了FReT和查询处理方法在不影响查询结果准确性的前提下，在较小的延迟和开销方面表现良好。