Abstract
Environmental pollutions caused by improper abandoned cartridges increase dramatically nowadays. In Hong Kong, due to abundant quantity of cartridges being used, producers have to optimise their forward and reverse networks to maximise the recycling rate and their profits. In this paper, a comprehensive closed-loop supply chain (CLSC) model is established. This model contains eight partners in CLSC and describes the existing cartridge recycling situation in Hong Kong. In the literatures, many CLSC models were established and studied, but few of them analysed the delivery activity for different kinds of materials extracted from the used products and also, few papers studied the situation that used products are classified into good and poor quality. In this model, delivery activities of different materials are considered and the used cartridges are classified into good-quality ones and poor-quality ones. Producers will have different methods to process them. This problem is formulated into an Integer programming model. Since both delivery routes and delivery quantities problems are known to be NP hard, a novel modified two-stage genetic algorithm (GA) is proposed. A new two-stage encoding algorithm in the proposed GA reinforces the genetic searching ability in tackling this kind of problem. As the model is new in literature, we used Integer Programming to solve the testing instances and benchmark with the proposed algorithm. The results show that a near-optimal solution can be obtained by the proposed GA in a much shorter computational time.
Original language | English |
---|---|
Pages (from-to) | 3120-3140 |
Number of pages | 21 |
Journal | International Journal of Production Research |
Volume | 53 |
Issue number | 10 |
DOIs | |
Publication status | Published - 1 Jan 2015 |
Keywords
- closed-loop supply chain
- genetic algorithm
- linear programming
- reverse distribution
ASJC Scopus subject areas
- Industrial and Manufacturing Engineering
- Management Science and Operations Research
- Strategy and Management