A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Enhancing the Performance of Multiobjective Evolutionary Algorithms for Sanitary Sewer Rehabilitation Problems
The application of evolutionary algorithms (EA) to optimize the rehabilitation of existing sanitary sewer systems is challenging because sewer network are complex and requires computationally demanding hydraulic models to obtain accurate representation of the system. Additionally, the large number of conduits in a typical sewer network makes it difficult to find the near optimal solutions within a few number of iterations of optimization algorithms. To address this problem, there is a need for EA operators that requires fewer number of function evaluations and converge to near optimum solutions faster. A new operator is explored to enhance the performance of multiobjective evolutionary algorithms (MOEA) for sanitary sewer rehabilitation optimization problem. The proposed operator is based on the nondominated sorting evolutionary strategies (NSES) which combines the Pareto optimality of the nondominated sorting genetic algorithm (NSGA II) with evolution strategies (ES). The operator is based on a graph of topologically connected conduits so as to guide the search toward known SSOs locations, thereby speeding up the convergence time. The MOEA is designed to find solutions that address two conflicting objectives: maximize sanitary sewer overflow (SSO) reduction and minimize rehabilitation cost. The hydraulics of the network is modeled using the EPA storm water management model (SWMM). The proposed operator is applied to an existing sewer network in the eastern San Antonio water system (SAWS) network.
Enhancing the Performance of Multiobjective Evolutionary Algorithms for Sanitary Sewer Rehabilitation Problems
The application of evolutionary algorithms (EA) to optimize the rehabilitation of existing sanitary sewer systems is challenging because sewer network are complex and requires computationally demanding hydraulic models to obtain accurate representation of the system. Additionally, the large number of conduits in a typical sewer network makes it difficult to find the near optimal solutions within a few number of iterations of optimization algorithms. To address this problem, there is a need for EA operators that requires fewer number of function evaluations and converge to near optimum solutions faster. A new operator is explored to enhance the performance of multiobjective evolutionary algorithms (MOEA) for sanitary sewer rehabilitation optimization problem. The proposed operator is based on the nondominated sorting evolutionary strategies (NSES) which combines the Pareto optimality of the nondominated sorting genetic algorithm (NSGA II) with evolution strategies (ES). The operator is based on a graph of topologically connected conduits so as to guide the search toward known SSOs locations, thereby speeding up the convergence time. The MOEA is designed to find solutions that address two conflicting objectives: maximize sanitary sewer overflow (SSO) reduction and minimize rehabilitation cost. The hydraulics of the network is modeled using the EPA storm water management model (SWMM). The proposed operator is applied to an existing sewer network in the eastern San Antonio water system (SAWS) network.
Enhancing the Performance of Multiobjective Evolutionary Algorithms for Sanitary Sewer Rehabilitation Problems
Ogidan, Olufunso (author) / Itaquy, Bruno (author) / Giacomoni, Marcio (author)
World Environmental and Water Resources Congress 2016 ; 2016 ; West Palm Beach, Florida
2016-05-16
Conference paper
Electronic Resource
English
| British Library Online Contents | 2017
Rehabilitation of sanitary sewer lines
| Engineering Index Backfile | 1966