StoDyM_MOEA

For a given number of check dams in a catchment, this numerical framework optimizes the dams' location and storage capacity in the catchment's stream path. Please take a look on the StoDyM inputs before read the following inputs required for the Multi-Objective Evolutionary Algorithm (MOEA).

Requirement

Python 3.6 and Platypus of Python 3, which is a framework for evolutionary computing in Python (https://github.com/Project-Platypus/Platypus)

Inputs

• cpn_d: Dictionary of catchments' pixel number
• cdn_d: Dictionary of catchments' dam number
• cdscul_d: Dictionary of upper limit of dams' initial total storage capacity
• Storage capaicty of dams: It can be provided by either Integer (dscdv_i = True) or Subset (dscdv_i = False) class from Platypus
  • cdsci_d: Dictionary of dams' storage capacity in terms of Integer class. The user needs to provide a multiplier of Integer (dsc_mult, default value is 1000) to get the actual dams' storage capacity. It can be adjusted according to the range of Integer and storage capacity. Here, the user cannot set a specific search space of dams' storage capacity.
  • cdscs_d: Dictionary of dams' storage capacity in terms of Subset class. Here, the user can set a specific search space of dams' storage capacity by providing a list (default is list(range(start, end, step))).
• eps_v: Value of epsilon required for EpsNSGAII (default is 0.001)
• cn_s: Name of catchment
• (o_no, ps_no, fe_no, s_no, cc_no): Numbers of objectives (o_no), population size (ps_no), function evaluations (fe_no), seed (s_no), and computer core (cc_no)
• od_l: objective decision list of maximization and minimization in case user change any objective
• cs_l: constraint list in case user change any constraint
• u_p: array of Utopian point (1 for maximization and 0 for minimization)

Note that each dictionary contains two keys "Shejiagou" and "Majiagou", since the framework is tested on these two catchments. The user can insert arbitraty number of catchments.

Outputs

For each combination of algorithm and problem, the numerical framework computes the Pareto-front set from the merged solutions of all seeds and saves the Pareto-front set (sheet_name="Pareto_Front") and simultion time (sheet_name="Time") in excel file. Each row of the "Pareto_Front" sheet represents a Pareo-front solution with the information of

• dams' location (DO, D1 etc.)
• dams' storage capacity (SCO, SC1 etc. in m³)
• intial total storage capacity (ITSC, in m³)
• intial total drainage area as a percentage of the catchment area (ITDA)
• four objective values of average life expectancy (O-LE), storgae dynamics (O-SD), short-term SDR (O-SDR,ST), and long-term SDR (O-SDR,LT)
• Life expectancy of each dam (LE0, LE1 etc. in years)
• Average life expectancy (LE_A) in years
• Distance from Utopian point (last column)

Note that SCN and LEN represent the features of DN dam, where N = 0, 1, 2, ..., n and n is the number of dams