Excel 模型的单目标多整数变量优化 - Python Pymoo 库

Question

我正在尝试使用 Pymoo 优化大型 Excel 模型，我已经设置了通过 pyWin32 与 Excel 进行通信的问题。

在 VSCode 中调试时出现以下异常：

Exception
('Problem Error: F can not be set, expected shape (100, 1) but provided (1, 1)', ValueError('cannot reshape array of size 1 into shape (100,1)'))
ValueError: cannot reshape array of size 1 into shape (100,1)

During handling of the above exception, another exception occurred:

  File "C:\Database\Python\RSG\RSG Opt.py", line 71, in <module>
    res = minimize(
          ^^^^^^^^^
Exception: ('Problem Error: F can not be set, expected shape (100, 1) but provided (1, 1)', ValueError('cannot reshape array of size 1 into shape (100,1)'))

我正在尝试优化单个目标，所以我不知道为什么它需要一个包含 100 个值的数组。我绝对不会将变量设置为 100。Pymoo 是否需要一个数组来保存第一个总体 100 的所有 100 个评估？

完全意识到我的代码可能存在其他问题，可能错误地设置了数组大小，因此完整的代码在这里：

import win32com.client as win32
import numpy as np
from pymoo.core.problem import Problem
from pymoo.algorithms.soo.nonconvex.ga import GA
from pymoo.optimize import minimize

# Connect to Excel
excel = win32.gencache.EnsureDispatch('Excel.Application')
excel.Visible = True  # Keep Excel hidden

# Reference the active workbook (assuming it's already open)
workbook = excel.ActiveWorkbook

# Prompt for the sheet name and range
sheet_name = input("Enter the sheet name: ")
range_address = input("Enter the variable range address (e.g., A1:B10): ")
target = input("Enter the objective cell address (e.g., A1:B10): ")

# Reference the specified sheet and range
try:
    worksheet = workbook.Sheets(sheet_name)
    variable_range = worksheet.Range(range_address)
    objective_cell = worksheet.Range(target)

except Exception as e:
    print(f"Error: {e}")
    excel.Quit()
    quit()

# Read the number of variables based on the number of rows in the range
num_vars = variable_range.Rows.Count
Vars=np.array(variable_range.Value)
trans_vars = np.transpose(Vars)

# Read the lower bounds from Excel (assuming they are in column -2)
lower_bounds_range = variable_range.GetOffset(0,-2)
lower_bounds = [int(cell.Value) for cell in lower_bounds_range]

# Read the upper bounds from Excel (assuming they are in column -1)
upper_bounds_range = variable_range.GetOffset(0, -1)
upper_bounds = [int(cell.Value) for cell in upper_bounds_range]


# Define the Optimization Problem
class ExcelOptimizationProblem(Problem):
    def __init__(self, num_vars, lower_bounds, upper_bounds):
        super().__init__(n_var=num_vars, n_obj=1, n_constr=1, xl=lower_bounds, xu=upper_bounds, type_var=int)

    def _evaluate(self, x, out, *args, **kwargs):
        # Set the decision variable values in Excel
        for i, val in enumerate(x):
            variable_range.Cells(1,1).GetOffset(i, 0).Value = val

        excel.Calculate()

        objective_value = float(objective_cell.Value)

        constraint_value = 0

        out["F"] = [-objective_value]
        out["G"] = [-constraint_value]

problem = ExcelOptimizationProblem(
    num_vars, lower_bounds, upper_bounds
    )

algorithm = GA(
    pop_size=100,
    eliminate_duplicates=True)

res = minimize(
    problem, algorithm, ('n_gen', 10), verbose=True
    )

print("Best solution found: \nX = %s\nF = %s" % (res.X, res.F))

workbook.Close()
excel.Quit()

有人可以给我一些指导吗？

提前致谢。

Answer 1

你的变量数量

n_vars

与上下限的长度一致吗？ 错误到底是在哪一行代码中产生的？