第三代无线网络规划新的优化模型及其可信智能优化方法研究

项目名称： 第三代无线网络规划新的优化模型及其可信智能优化方法研究

项目编号： No.60873099

项目类型： 面上项目

立项/批准年度： 2009

项目学科： 轻工业、手工业

项目作者： 王宇平

作者单位： 西安电子科技大学

项目金额： 30万元

中文摘要： 无线网络规划问题可以分为多个子规划问题，这些问题可抽象成多种类型的优化模型。对于网络规划中的任务调度问题，在考虑了任务执行时的安全性和可靠性等因素的前提下，我们建立了几类新的网络任务调度模型，这些模型是目前为止对这类问题提出的极少数融合安全性和可靠性的优化模型。对于网络规划中的优化模型为确定性单目标优化问题时，设计了几种能够快速收敛到全局最优解的智能优化方法，这些方法克服了类似算法存在的早熟、精度低、收敛速度慢等缺陷。当优化模型为确定性多目标优化问题时，设计了几种有效的多目标进化算法，这些算法能够求出分布均匀且广泛的Pareto最优解，特别是对于带约束的多目标优化问题，克服传统约束处理方法计算量大且易产生计算溢出的缺陷。当优化模型为动态优化问题时，设计了几种能够快速收敛到全局最优解的智能优化方法，这些算法克服了预测复杂和所需数据较多的缺陷，能够快速准确的跟踪最优解的变化。对于优化模型中的非线性双层规划模型，设计了几种有效的且可保证全局收敛的进化算法，为网络规划提供了一种新的有效的优化工具。大量的计算机实验表明了这些优化模型和智能算法是可信的。

中文关键词： 网络规划;网络优化模型;智能优化方法

英文摘要： Some optimization models are set up and considered for radio network planning problems, including single objective and multiobjective optimization models, task scheduling model,etc. For the task scheduling problems in the network planning, the models considering the security and reliability during the tasks execution are few and we established some new such models. For the deterministic single-objective optimization models in network planning, several new rapidly convergent intelligent optimization methods are proposed. These methods overcome the drawbacks of premature, low precision and slow convergence of the existing methods. For the deterministic multi-objective optimization models, we design several effective multi-objecive evolutionary algorithms, which can find uniformly distributed and wide spread Pareto optimal solutions. Especially, for the constrained multi-objective optimization problems, the proposed algorithms overcome the drawbacks of large computation and easy overflow of the similar constraint handling methods. For the uncertainty optimization models, we design several effective dynamical evolutionary algorithms, which can converge to the global optimum rapidly. These algorithms overcome the drawbacks of the complex prediction and requirement of large information, and can track the variation of optimum rapidly and accurately. For the nonlinear bilevel optimization model in the network planning, several effective and convergence-guaranteed evolutionary algorithms are designed. The research of this project provides a new kind optimization tool for network planning problems.

英文关键词： Network planning; network optimization model; intelligent optimization method