Enabling multiple users to transmit on the same channel, simultaneously, is an essential part of the future generation of communication systems. Optimization of transmit power has a significant impact on the performance of communication systems. Under spectrum re-use, the transmit power of one user causes interference to the other users on the shared channels, also known as interference channels (ICs). Under IC model, the structure of power allocation problems becomes complex. Recently, artificial intelligence (AI) techniques have shown promising results in various wireless communication problems. This paper aims to explore various AI techniques to optimize power allocation in multi-user IC transmission. Sum rate maximization problem subject to guaranteed per user performance under limited available power budget at each transmitting node is considered. Specifically, we first investigate the performance of Particle Filter and Particle Swarm Optimization based search algorithm. Then, Linear Regression and Neural Network based predictors are designed to reduce the overall time complexity. Finally, extensive simulations are presented to compare the performance of obtained solutions.

使多个用户同时在同一信道上传输是下一代通信系统的重要组成部分。发射功率的优化对通信系统的性能有重大影响。在频谱复用下，一个用户的发射功率会在共享信道（也称为干扰信道（IC））上对其他用户造成干扰。在IC模型下，功率分配问题的结构变得复杂。近年来，人工智能（AI）技术已在各种无线通信问题中显示出令人鼓舞的结果。本文旨在探索各种AI技术，以优化多用户IC传输中的功率分配。考虑在每个发射节点的有限可用功率预算下，要保证每个用户性能的求和率最大化问题。具体来说，我们首先研究基于粒子滤波和粒子群优化的搜索算法的性能。然后，设计基于线性回归和神经网络的预测器以减少总体时间复杂度。最后，提出了广泛的仿真以比较所获得解决方案的性能。