CAJ | 학술논문

根据我国目前在招标投标过程中评标方式存在的缺陷，运用组合优化思想，将经过粒子群算法优化过的径向基神经网络模型（PSO-RBF）运用到招标评标的实践中。该模型作为一种新型的评标方法，先是利用粒子群优化算法对单纯的径向基神经网络参数进行优化，进而训练和测试优化后的RBF神经网络。通过对比分析单纯的RBF神经网络和经过参数优化后的RBF神经网络模型，结果表明后者比前者在性能和效率上更加优越，在招标投标系统中更加合理、科学。最后指出了该模型还存在的缺陷有待进一步的研究。
근거아국목전재초표투표과정중평표방식존재적결함，운용조합우화사상，장경과입자군산법우화과적경향기신경망락모형（PSO-RBF）운용도초표평표적실천중。해모형작위일충신형적평표방법，선시이용입자군우화산법대단순적경향기신경망락삼수진행우화，진이훈련화측시우화후적RBF신경망락。통과대비분석단순적RBF신경망락화경과삼수우화후적RBF신경망락모형，결과표명후자비전자재성능화효솔상경가우월，재초표투표계통중경가합리、과학。최후지출료해모형환존재적결함유대진일보적연구。
For the shortcoming of evaluation methods of bidding system in china,the radical basis function based on particle swarm optimization(PSO-RBF)algorithm is applied to biding practice. This model is a new bid evaluation method. firstly,radical basis function(RBF)is optimized by particle swarm optimization(PSO). Secondly,the optimized RBF neural network is trained and tested,and then used to predict a sample. The paper compares the two models of RBF and PSO-RBF,and finds that the PSO-RBF model has better performance than RBF in the aspect of function and efficiency. The PSO-RBF model is more suitable for bidding system due to its reasonability. At last,the paper points out the limitation of the PSO-RBF model and directs path for further study.