APPROACHES TO MULTI-OBJECTIVE MULTI-TIER OPTIMIZATION IN URBAN DRAINAGE PLANNING

摘要

Flooding in urban areas was and is a matter of concern of governments in many countries. One of the possible solutions is to optimize the drainage system networks; funds allocated for this are however limited, so the problem should be posed as a constrained optimization problem (due to limited resources) in a multi-criterial context (due to many stakeholders with conflicting interests). A "multi-tier" approach is introduced, where different tiers correspond to different dimensions (or sub-problems) of the whole problem. Two multi-objective evolutionary algorithms are tested, NSGA-Ⅱ and ε-MOEA in solving two problems of drainage network rehabilitation. Both algorithms were evaluated by the indices traditionally used in multi-objective optimization. NSGA-Ⅱ's performance is marginally better, however ε-MOEA has other advantages.