On the effect of the selection of suppliers on the design of formulated products

An extended pooling problem is developed for the design of detergents, that include process and product design together with suppliers' selection. It is a multiperiod and multi-objective optimization problem since it considers the economic benefit and environmental impact as well as the contract length. This problem can be considered as a previous step to the integrated product, process and supply chain design. This type of problems is non-linear and non-convex. Two formulations are developed to tackle this problem. An MINLP and a reformulated NLP using a decision vector avoiding the use of binary variables. The NLP shows better computational performance in spite of the larger problem size. Furthermore, it is demonstrated that when considering different ingredients, formulations, pricing policies and suppliers, their selection can be adjusted until a reduction of almost 40% of CO2 emissions is achieved without the benefit decreasing more than 1.5%. (C) 2020 Elsevier Ltd. All rights reserved.