Dynamic Multi-objective Optimization Of Chemical Process Using Bare-bones MOPSO Algorithm

Chemical processes strictly are dynamic processes and usually need to optimize multiple targets. Dynamic multi-objective optimization is a complex and difficult research topic of process systems engineering. It is hard to serve effective decisions according to different conditions by simply converting to a single objective optimization problem. How to effectively solve this kind of problem has caught much attention from academia and industry.In this paper, a novel MOPSO algorithm is proposed that adopted the adaptive sampling distribution strategy, circular crowded sorting approach and new mutation operator to enhance the exploratory capability and the uniformity of distribution. It is proven that the new algorithm shows better performance compared with other proposed multi-objective optimizations with the result of solving three optimization problems. To solve constrained dynamic multi-objective optimization problems, a new kind of constraint handling method which uses double external archives to keep non-dominated solutions is proposed. And at the same time local search and hybrid mutation operator are exceeded. By combining the above mentioned strategy of Bare-bones MOPSO and control vector parameterization, an approach is proposed to solve the constrained dynamic optimization problems. The advantageous performance of Bare-bones MOPSO is validated by comparisons with NSGA-Ⅱ and a newly proposed SADE-aCD algorithm over eight test problems and two dynamic optimization problems.At the last of this paper a constrained multi-objective optimization model with the maximum ethylene production and minimum ethane production objectives for acetylene hydrogenation optimization problem is proposed. And with the use of the proposed algorithm, ethylene production is improved and at the same time the ethane production is less.