论文部分内容阅读
为了改善发酵过程补料优化控制的性能指标,提出一种基于改进的分解多目标粒子群优化算法的发酵过程补料优化控制方法。将发酵过程的多目标优化问题分解为若干单目标优化问题,对每个单目标优化问题应用2个粒子群协同搜索最优解,其中一个种群对连接向量进行搜索,改进了连接向量的更新方式,另一个种群将决策空间分解,使用不同种群对不同维度的解向量进行优化,并在进化过程中使用前一个种群的最优解构成完整的解向量评价每个子群的最优解,提高了最优解评价的准确性,保证算法收敛到全局最优解。将该算法用到在工业酵母发酵过程模型中,对补料速率进行优化控制,并同时优化2个性能指标,即菌体浓度最大化和乙醇浓度最小化,并与基本的分解粒子群优化算法的优化结果作对比,该算法获得的Pareto前沿在基本分解粒子群优化算法获得的Pareto前沿下方,并且分布更加完整。实验证明该算法能够获得最优补料速率,使工业酵母发酵过程的菌体浓度提高了约9%,乙醇浓度降低了约为15%,为发酵过程补料优化控制提供了一种有效方法。
In order to improve the performance of feeding process control in fermentation process, a feed optimization control method based on improved decomposition multi-objective particle swarm optimization algorithm was proposed. The multi-objective optimization problem of fermentation process is decomposed into several single-objective optimization problems. For each single-objective optimization problem, two particle swarms cooperatively search for the optimal solution. One of the swarms searches for the connection vector and improves the updating method of the connection vector , The other population decomposes the decision space and uses different populations to optimize the solution vectors of different dimensions. In the evolution process, the optimal solution of the previous population is used to construct a complete solution vector to evaluate the optimal solution of each sub-population and improve The accuracy of the optimal solution evaluation ensures that the algorithm converges to the global optimal solution. The algorithm was used in the industrial yeast fermentation process model to optimize the feed rate control, and at the same time to optimize two performance indicators, namely, the maximum concentration of bacteria and ethanol concentration, and with the basic decomposition of particle swarm optimization algorithm , The Pareto front obtained by this algorithm is under the Pareto front obtained by the particle swarm optimization algorithm with basic decomposition, and the distribution is more complete. Experiments show that the algorithm can get the optimal feeding rate, which can increase the concentration of the yeast in the fermentation process by about 9% and the ethanol concentration by about 15%. It provides an effective method for the feedstuff optimization in the fermentation process.