Every year about one third of the food production intended for humans gets lost or wasted. This wastefulness of resources leads to the emission of unnecessary greenhouse gas, contributing to global warming and climate change. The solution proposed by the SORT project is to “recycle” the surplus of food by reconditioning it into animal feed or fuel for biogas/biomass power plants. In order to maximize the earnings and minimize the costs, several choices must be made during the reconditioning process. Given the extremely complex nature of the process, Decision Support Systems (DSSs) could be helpful to reduce the human effort in decision making. In this paper, we present a DSS for food recycling developed using two approaches for finding the optimal solution: one based on Binary Linear Programming (BLP) and the other based on Answer Set Programming (ASP), which outperform our previous approach based on Constraint Logic Programming (CLP) on Finite Domains (CLP(FD)). In particular, the BLP and the CLP(FD) approaches are developed in ECLiPSe, a Prolog system that interfaces with various state-of-the-art Mathematical and Constraint Programming solvers. The ASP approach, instead, is developed in clingo. The three approaches are compared on several synthetic datasets that simulate the operative conditions of the DSS.
Declarative and Mathematical Programming approaches to Decision Support Systems for food recycling / Chesani, F.; Cota, G.; Gavanelli, M.; Lamma, E.; Mello, P.; Riguzzi, F.. - In: ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE. - ISSN 0952-1976. - 95(2020), pp. 1-11. [10.1016/j.engappai.2020.103861]
|Appare nelle tipologie:||1.1 Articolo su rivista|