ABSTRACT

We consider a conflict-free minimum latency data aggregation problem that occurs in different wireless networks. Given a network that is presented as an undirected graph with one selected vertex (a sink), the goal is to find a spanning aggregation tree rooted in the sink and to define a conflict-free aggregation minimum length schedule along the arcs of the tree directed to the sink. Herewith, at the same time slot, each element of the network can either send or receive at most one message. Only one message should be sent by each network element during the whole aggregation session, and the conflicts caused by signal interference should be excluded. This problem is NP-hard and remains NP-hard even in the case when the aggregation tree is given. Therefore, the development of efficient approximation algorithms is very important for this problem. In this paper, we present new heuristic algorithms based on genetic local search and variable neighbourhood search metaheuristics. We conducted an extensive simulation that demonstrates the superiority of our algorithms compared with the best of the previous approaches.

摘要

我们考虑发生在不同无线网络中的无冲突最小延迟数据聚合问题。给定一个网络，该网络呈现为具有一个选定顶点（汇）的无向图，目标是找到一个以汇为根的生成聚合树，并沿有向树的弧定义一个无冲突的聚合最小长度调度到水槽。因此，在同一时隙，网络的每个元素最多可以发送或接收一个消息。在整个聚合会话期间，每个网元只应发送一条消息，应排除由信号干扰引起的冲突。这个问题是 NP-hard 并且即使在给出聚合树的情况下仍然是 NP-hard。因此，开发有效的逼近算法对于这个问题非常重要。在本文中，我们提出了基于遗传局部搜索和变量邻域搜索元启发式的新启发式算法。我们进行了广泛的模拟，证明了我们的算法与以前最好的方法相比的优越性。