Efficient well placement optimization coupling hybrid objective function with particle swarm optimization algorithm

2020 ◽  
Vol 95 ◽  
pp. 106511 ◽  
Author(s):  
Shuaiwei Ding ◽  
Ranran Lu ◽  
Yi Xi ◽  
Guangwei Liu ◽  
Jinfeng Ma
Author(s):  
Ziyang Li ◽  
Quan Zhou ◽  
Yunfan Zhang ◽  
Ji Li ◽  
Hongming Xu

The self-adaptive and highly robust proportional-integral-like fuzzy knowledge–based controller has been developed to regulate air–fuel ratio for gasoline direct injection engines, in order to improve the transient response behaviour and reduce the effort to be spent on calibration of parameter settings. However, even though the proportional-integral-like fuzzy knowledge–based controller can automatically correct the initially calibrated proportional and integral parameters, a more appropriate selection of controller parameter settings will lead to better transient performance. Thus, this article proposes an enhanced intelligent proportional-integral-like fuzzy knowledge–based controller using chaos-enhanced accelerated particle swarm optimization algorithm to automatically define the most optimal parameter settings. An alternative time-domain objective function is applied for the transient calibration programme without the need for prior selection of the search-domain. The real-time transient performance of the enhanced controller is investigated on the air–fuel ratio control system of a gasoline direct injection engine. The experimental results show that the enhanced proportional-integral-like fuzzy knowledge–based controller based on chaos-enhanced accelerated particle swarm optimization is able to damp out the oscillations with less settling time (up to 75% reduction) and less integral of absolute error (up to 64.07% reduction) compared with the conventional self-adaptive proportional-integral-like fuzzy knowledge–based controller. Repeatability tests indicate that the chaos-enhanced accelerated particle swarm optimization algorithm–based proportional-integral-like fuzzy knowledge–based controller is also able to reduce the mean value of objective function by up to 10.61% reduction and the standard deviation of the objective function by up to 28.29% reduction, compared with the conventional accelerated particle swarm optimization algorithm–based proportional-integral-like fuzzy knowledge–based controller.


Particle Swarm Optimization, a nature based stochastic evolutionary algorithm that iteratively tries to improvise the solution pertaining to a particular objective function. The problem becomes challenging if the objective function is not properly identified nor it is properly been evaluated which results in slow convergence and inability to find the optimal solution. Hence, we propose a novel rough set based particle swarm optimization algorithm using golden ratio principle for an efficient feature selection process that focusses on two objectives: First, that results in a reduced subset of features without conceding the originality of the data and the second is that yields a high optimal result. Since many subset of features might result with a meaningful solution, we have used the golden ratio principle to extract the most reduced subset with a high optimal solution. The algorithm has been tested over several benchmark datasets. The results shows that the proposed algorithm identifies a small set of features without convincing the optimal solution, thus able to achieve the stated objectives.


Sign in / Sign up

Export Citation Format

Share Document