Improvements to OSTRICH-SWMM: A multi-LID and multi-algorithm optimization software tool

Maria Torres, Alan Rabideau, Shawn Matott and Zhenduo Zhu

ABSTRACT

OSTRICH-SWMM, developed at the University at Buffalo, is a public domain Python module that connects EPA SWMM (Storm Water Management Model) with OSTRICH (Optimization Software Toolkit for Research Involving Computational Heuristics). Since its initial release in 2019, OSTRICH-SWMM has been used in several studies involving the optimal design of green infrastructure (GI). These studies have considered both neighborhood-scale and city-wide GI implementations (i.e., rain barrels, cisterns, porous pavements, and green roofs) along with numerical experiments that explored alternative compositions of parallelized workloads in high-performance computing (HPC) environments. The results of these studies have motivated several important updates to the OSTRICH-SWMM software, including: (1) migrating to Python 3.x, (2) adding support for the entire set of LID features currently modeled by SWMM, (3) improving optimizer behavior via novel “seeding” strategies; and (4) efficiently optimizing the number of LID installations within a set of pre-determined locations. These improvements have been evaluated using a carefully constructed suite of test problems derived from CSOsheds in City of Buffalo, New York. 


Permanent link: