INTEGRATED CONSTRAINT LOGIC AND MATHEMATICAL

C. VAN CAUWENVERGHE, A. BANDONI

Rio de Janeiro, Brasil

Congreso; ENPROMER 2005; 2005

Mathematical Programming (MP) and Constraint Logic Programming (CLP) are two complementary approaches for solving complex combinatorial optimisation problems. CLP essentially provides an environment in which one can describe and solve a mixture of algebraic and logical constraints. Nowadays, it is applied mostly to planning and scheduling constraints, for solving various real-world industrial applications. These types of problems are also in the domain of MP, and a comparison between both approaches is therefore meaningful. Recent developments have shown that using the two approaches together can be quite beneficial, both in terms of solution time and expressiveness of the language. Many problems can also be formulated as a Constraint Satisfaction Problem (CSP), although recognize them could not be easy, and consequently, we can fail to make use of specialised techniques for solving them. Its role in logic programming has formerly been recognised in Artificial Intelligence, but in recent years, CSP has come to be seen as a key mathematical tool in many applications, such as temporal reasoning, resource allocation, scheduling, amongst others. This paper presents a combined strategy of CLP and MP to solve planning and scheduling problems in the food industry. The approach is applied to a typical apple farm from southern Argentina. Results from both individual and combined optimisation strategies are presented. Mathematical Programming (MP) and Constraint Logic Programming (CLP) are two complementary approaches for solving complex combinatorial optimisation problems. CLP essentially provides an environment in which one can describe and solve a mixture of algebraic and logical constraints. Nowadays, it is applied mostly to planning and scheduling constraints, for solving various real-world industrial applications. These types of problems are also in the domain of MP, and a comparison between both approaches is therefore meaningful. Recent developments have shown that using the two approaches together can be quite beneficial, both in terms of solution time and expressiveness of the language. Many problems can also be formulated as a Constraint Satisfaction Problem (CSP), although recognize them could not be easy, and consequently, we can fail to make use of specialised techniques for solving them. Its role in logic programming has formerly been recognised in Artificial Intelligence, but in recent years, CSP has come to be seen as a key mathematical tool in many applications, such as temporal reasoning, resource allocation, scheduling, amongst others. This paper presents a combined strategy of CLP and MP to solve planning and scheduling problems in the food industry. The approach is applied to a typical apple farm from southern Argentina. Results from both individual and combined optimisation strategies are presented.