Application Research Of Optimization Design For Drainage Network By Ant Colony Algorithm

Drainage network is one of vital components of urban municipal infrastructures.It accounts for70%~80%proportion in the drainage system. So it’s very important ineconomic significance of optimization design and the engineering investment costreducing.In the past, many researchers excessively depended on the programming to selectfeasible variable sets based on hydraulic calculation formulas after choosing thedecision variables. The connection of hydraulic parameters and how to select feasiblevariable sets were lack of study. Besides, the computer programming compilation wasentangled. The points mentioned above are hindering the development of optimizationdesign of drainage network. Ant colony algorithm (ACG) has been showed greatsuperiority on handling discrete optimization problems. It is necessary and practicableto apply for the optimization design on drainage network.In this research, first, the feasible variables space is artificially set up throughanalysis of hydraulic constraint conditions in order to simplify the mathematicalmodel limited conditions and the computer programming compilation. And then, thegoal would be achieved based on the modified ACG and MATLAB with obviousadvantage on settling matrix problems. Finally, the optimization effect is tested by theexamples.This research focuses on the optimization design of the decided pipeline. Themethod takes diameters and fullness degree of the pipeline as the decision variable.Two-dimensional feasible variables space meeting the requirements of design code isestablished through the analysis of hydraulic constraint conditions and monotonicityof hydraulic parameters. Considering the corresponding nonlinear programmingmixed discrete variables issue, discretization is utilized to turn feasible variable spaceinto feasible variable set. The mathematical model of drainage network is set up basedon the minimum cost of total investment.Essential hydraulic parameters of ACG are valued in proportion mode to reducethe influences of problem scale. Feasibility, effectiveness and high efficiency areverified by the examples.