Plant-wide optimal waste management

Abstract Synthesis of optimal waste treatment policies on a plant wide level is a formidable challenge due to (i) the knowledge necessary to do selection of treatment steps and (ii) the combinatorial complexity of the resulting problem space. This paper presents a hybrid methodology for an automatic synthesis of feasible alternatives, the superstructure generation, followed by rigorous combinatorial optimization. In step one, evaluation of waste properties against regulatory limits as well as a relaxed set of technology selection criteria gives rise to a tree of feasible treatment options. In step two, rigorous integer programming techniques expose optimal waste management policies with best trade-off between different levels of environmental, ecological and logistic constraints. Results of the application of the methodology onto a case study from the pharmaceutical industry are discussed. These figures demonstrate clearly the huge degree of flexibility that can be obtained by systematic waste management using computerized methods.