混合编码免疫算法在船舶载重计量的应用

提出一种混合编码免疫辨识算法, 并应用于船舶载重货物计量的模型辨识. 该算法结合免疫算法优越的全局搜索性能与GP算法简洁的结构树编码方法, 将抗体编码为模型结构树编码与模型参数编码的组合, 通过对抗体结构与参数的免疫操作, 实现全局寻优及非线性模型的结构与参数的一体化辨识. 算法不依赖对象的先验知识, 辨识的模型结构简单、易于理解. 仿真及船舶载模型的辨识应用验证了本算法的有效性及较强的非线性逼近能力.

Application of combinatorial-coded artificial immune system(AIS) to ship cargo weighing

A combinatorial-coded artificial immune system(AIS) is presented and applied to the modeling of ship cargo weighing. This algorithm possesses the advantage of immune algorithm in global searching and the conciseness of genetic program(GP) in structure-tree. It codes the antibodies in a combination of structure-tree-coding model and parametercoding model. By applying immune operations to both structures and parameters of antibodies, the global optimization and the identification of both structures and parameters of the nonlinear system are thus realized. The algorithm is independenct of prior knowledge; and the identified model is of simple structure to be understood. Simulations and applications of the identification to the ship cargo weighing model demonstrate the efficacy and efficiency of the proposed algorithm.