摘要
目前,遗传算法作为一种基于人工智能技术的优化算法在电力系统的电源规划中已经得到广泛应用,然而其遗传操作繁杂、计算量大,易产生退化、早熟等问题使其应用受到局限。文章提出一种基于自然分段式编码的改进免疫算法(Modified Immune Algorithm,MIA),使用该方法可以大大减少电源规划的计算量,提高计算效率,并可充分利用特征信息灵活求解。该算法具有全局多峰搜索能力以及优异的收敛特性,同时可避免退化、早熟等问题的发生。算例结果表明,该算法可以较好地求解电源规划问题,并且具有广阔的发展空间。
At present as an optimized algorithm based on artificial intelligence, genetic algorithm is widely used in generation expansion planning. However, due to complicated operation, huge calculating, degeneration and immature convergence the utilization of genetic algorithm in practice is limited. Here, a modified immune algorithm (MIA) based on divided natural coding is proposed by which the calculation amount of generation expansion planning can be considerably reduced and the computational efficiency can also be improved, the solution of the planning can be flexibly achieved by character information. The proposed algorithm possesses global multiamplitude search ability and excellent convergence, so the phenomena of degeneration and paedogenesis can be avoided. The results of calculation examples show that this algorithm can be used to solve the generation expansion planning and it can be further developed too.
出处
《电网技术》
EI
CSCD
北大核心
2004年第11期38-44,共7页
Power System Technology
关键词
电力系统
电源规划
改进免疫算法
人工智能
电力负荷
负荷预测
Computational methods
Convergence of numerical methods
Genetic algorithms
Immunology
Optimization
Planning