7. Conclusion
By using a multiple echelon network, it is shown that the proposed MINLP model can be used to represent complex reverse logistics processes in handling product recovery and remanufacturing. The objective of the mathematical model is profit maximization with consideration of multi-product and multi-module returnable products and a variety of recycling channels for the returned products. The numerical experiments reflect the real recovery processes of the used bulk waste products in a city of Taiwan, and the results show some distinctive features.
The proposed model is a generic model and can represent current reverse logistics operated by some industries using existing distribution centers, dismantling centers, warehouses, and factories for returned products. Facility location and available capacity are important issues in reverse logistics networks. By identifying the critical activities and related requirements involved in the processes of reverse logistics operations, the proposed model can determine the optimization of facility locations, their state of operation (open or closed), capacity utilization and the optimal flow of returned products and dismantled modules in the reverse network. The designed model is validated and tested through the proposed hybrid GA by using a real-life example of recycling bulk waste in Taoyuan City, Taiwan. Also, the post-optimality analysis and comparison show the proposed model performs better than current reverse operations in the city.
