Keywords

multi-objective optimization; life cycle assessment; constraint programming; mixed-integer programming; expensive optimization, computational sustainability.

Location

Session D4: Water Resource Management and Planning - Modelling and Software for Improving Decisions and Engaging Stakeholders

Start Date

12-7-2016 10:30 AM

End Date

12-7-2016 10:50 AM

Description

Many real-world multi-objective optimization (MOO) problems rely on computationally expensive simulators of industrial processes and require solutions within a limited time budget. In this context, we propose a heuristic approach which aims at building a surrogate problem model, solvable by computationally efficient optimization methods, in order to quickly provide a sufficiently accurate estimation of the Pareto front. The proposed approach generates a multi-objective mixed-integer programming (MO-MIP) proxy model of the MOO problem using sensitivity-based piece-wise linear approximation of objectives and constraints. The approximation of the Pareto front is obtained by applying the ε-constraint method to the multi-objective surrogate problem, transforming it into a desired number of single objective (SO) MIP problems. The paper further ex- plores the pros and cons of three algorithms for the solution of the SO-MIP problems namely constraint programming (CP), MIP, and constraint integer programming (CIP) which integrates CP and MIP methods. In the context of computational sustainability, the proposed methodology is successfully applied to the cost versus life cycle assessment (LCA)-based environmental optimization of potable water production plants (PWPPs). The numerical results obtained indicate that the proposed approach converges much faster to the Pareto front than the state-of-the-art metaheuristic algorithm SPEA2.

 
Jul 12th, 10:30 AM Jul 12th, 10:50 AM

Constraint Programming versus MIP for LCA-based Multi-Objective Optimization of Sustainable Potable Water Production Plants

Session D4: Water Resource Management and Planning - Modelling and Software for Improving Decisions and Engaging Stakeholders

Many real-world multi-objective optimization (MOO) problems rely on computationally expensive simulators of industrial processes and require solutions within a limited time budget. In this context, we propose a heuristic approach which aims at building a surrogate problem model, solvable by computationally efficient optimization methods, in order to quickly provide a sufficiently accurate estimation of the Pareto front. The proposed approach generates a multi-objective mixed-integer programming (MO-MIP) proxy model of the MOO problem using sensitivity-based piece-wise linear approximation of objectives and constraints. The approximation of the Pareto front is obtained by applying the ε-constraint method to the multi-objective surrogate problem, transforming it into a desired number of single objective (SO) MIP problems. The paper further ex- plores the pros and cons of three algorithms for the solution of the SO-MIP problems namely constraint programming (CP), MIP, and constraint integer programming (CIP) which integrates CP and MIP methods. In the context of computational sustainability, the proposed methodology is successfully applied to the cost versus life cycle assessment (LCA)-based environmental optimization of potable water production plants (PWPPs). The numerical results obtained indicate that the proposed approach converges much faster to the Pareto front than the state-of-the-art metaheuristic algorithm SPEA2.