Dynamic Optimization in Chemical Processes Using Region Reduction Strategy and Control Vector Parameterization with an Ant Colony Optimization Algorithm

Abstract

Two different approaches of the dynamic optimization for chemical process control engineering applications are presented. The first approach is based on discretizing both the control region and the time interval. This method, known as the Region Reduction Strategy (RRS), employs the previous solution in its next iteration to obtain more accurate results. Moreover, the procedure will continue unless the control region becomes smaller than a prescribed value. The second approach is called Control Vector Parameterization (CVP) and appears to have a large number of advantages. In this approach, control action is generated in feedback form, i.e., a set of trial functions of the state variables are expanded by multiplying by some unknown coefficients. By utilizing an optimization method, these coefficients are calculated. The Ant Colony Optimization (ACO) algorithm is employed as an optimization method in both approaches.

Control Vector Parameterization‐Based Adaptive Invasive Weed Optimization for Dynamic Processes

“…Rajesh et al reached a value of 0.76237 using control profile approximation under the frame of ACO . Asgari and Mahmoud obtained a value of 0.76160 employing CVP with the ACO algorithm by embedding a region reduction strategy . Qian et al got a value of 0.761595, which is the best literature result close to the analytical result using the CVP method with an iterative genetic algorithm .…”

“…This case is a benchmark dynamic optimization problem, which involves two state variables. The case has an analytical solution and the global optimum, thus, many researchers tested it as illustration of their approaches , , . The mathematical formulation of the case is stated as follows: …”

Harmony search algorithm with differential evolution based control parameter co-evolution and its application in chemical process dynamic optimization

“…ASGARI and PISHVAIE [24] obtained the values of 0.122 and 0.1599, while ZHANG et al [21] and RAJESH et al [23] reported 0.1235 and 0.12904, respectively.…”

Comparison of two Model Predictive Control for the ethanol production optimization of a two-stage bioreactor with cellular recycling

“…Place m ants (m = N x p) in each node. The algorithm proposed here follows the method of Zhang et al (2005) which placed the ants in each node at the difference of Asgari and Pishvaie (2008) which placed them in the nodes of the first column of time.…”