多目标优化黄金组合——Python熵权-模糊综合评价

import osimport timeimport numpy as npimport pandas as pdimport matplotlib.pyplot as pltimport seaborn as snsfrom matplotlib.colors import LinearSegmentedColormapINPUT_PATH = r"E:\DataAnalysis\Data.xlsx"SAVE_DIR = r"E:\DataAnalysis\EWM-FCE"os.makedirs(SAVE_DIR, exist_ok=True)os.makedirs(os.path.join(SAVE_DIR, "plots"), exist_ok=True)

# 粉色系渐变（薄雾粉→晶石紫）pink_palette = ["#F9EDF4", "#EDC9DD", "#E0A5C7", "#D180B1", "#C15B9B", "#B02F86"]custom_pink_cmap = LinearSegmentedColormap.from_list("custom_pink", pink_palette, N=256)# 森林绿单色渐变（森林绿→晨雾白绿）green_palette = ["#D4E7CF", "#B8D8B0", "#9CC891", "#7FB873", "#61A855"]custom_green_cmap = LinearSegmentedColormap.from_list("custom_green", green_palette, N=256)# 双色渐变（森林绿→浆果红）dual_palette = ["#61A855", "#7FB873", "#9CC891", "#B8D8B0", "#D4E7CF",                "#FFD6D4", "#FFBAB8", "#FE9E9D", "#F98082", "#F36069"]custom_dual_cmap = LinearSegmentedColormap.from_list("custom_dual", dual_palette, N=256)

plt.rcParams.update({    "font.family": ["Arial", "SimHei"],    "font.size": 10,    "axes.titlesize": 12,    "axes.labelsize": 11,    "xtick.labelsize": 9,    "ytick.labelsize": 9,    "legend.fontsize": 9,    "figure.dpi": 300,    "savefig.dpi": 600,    "savefig.bbox": "tight",    "axes.spines.top": False,    "axes.spines.right": False,    "grid.alpha": 0.3,    "grid.linestyle": "--",    "axes.unicode_minus": False})

df = pd.read_excel(INPUT_PATH, sheet_name="Sheet1")df_mean = df.groupby(["Treatment", "Factor1", "Factor2"]).mean().reset_index()group_cols = ["Treatment", "Factor1", "Factor2"]indicators = [col for col in df_mean.columns if col not in group_cols]

def data_preprocessing(df_mean, indicators, is_positive=None):    if is_positive is None:        is_positive = {ind: True for ind in indicators}    data_pos = df_mean[indicators].copy()    for ind in indicators:        if not is_positive[ind]:            data_pos[ind] = data_pos[ind].max() - data_pos[ind]    data_norm = pd.DataFrame()    for ind in indicators:        col_min = data_pos[ind].min()        col_max = data_pos[ind].max()        if col_max == col_min:            data_norm[ind] = 0.5   # 全列相同值，统一设为0.5，避免全0        else:            data_norm[ind] = (data_pos[ind] - col_min) / (col_max - col_min + 1e-8)    return data_pos, data_norm

def ewm_calculate_weights(data_norm):    col_sum = data_norm.sum(axis=0)    col_sum[col_sum == 0] = 1e-8    p_matrix = data_norm / col_sum    n_samples = data_norm.shape[0]    k = 1 / np.log(n_samples + 1e-8)    entropy = -k * (p_matrix * np.log(p_matrix + 1e-8)).sum(axis=0)    weights = (1 - entropy) / ((1 - entropy).sum() + 1e-8)    return p_matrix, entropy, weights

def fce_linear_weight(data_norm, weights):    fce_scores = np.dot(data_norm, weights.T)    return fce_scoresfce_scores = fce_linear_weight(data_norm, weights)result_df = df_mean[group_cols].copy()result_df["综合得分"] = np.round(fce_scores, 4)result_df["排名"] = result_df["综合得分"].rank(ascending=False, method="min").astype(int)

timestamp = time.strftime("%Y%m%d_%H%M%S")excel_path = os.path.join(SAVE_DIR, f"EWM-FCE_计算全过程_{timestamp}.xlsx")try:    with pd.ExcelWriter(excel_path, engine="openpyxl") as writer:        df_mean.to_excel(writer, sheet_name="1_原始均值数据", index=False)        data_pos_with_group = df_mean[group_cols].join(data_pos)        data_pos_with_group.to_excel(writer, sheet_name="2_正向化数据", index=False)        data_norm_with_group = df_mean[group_cols].join(data_norm)        data_norm_with_group.to_excel(writer, sheet_name="3_标准化数据", index=False)        p_with_group = df_mean[group_cols].join(pd.DataFrame(p_matrix, columns=indicators))        p_with_group.to_excel(writer, sheet_name="4_指标比重矩阵", index=False)        weight_df = pd.DataFrame({            "指标名称": indicators,            "信息熵": np.round(entropy, 4),            "熵权权重": np.round(weights, 4),            "权重占比(%)": np.round(weights * 100, 2)        }).sort_values("熵权权重", ascending=False)        weight_df.to_excel(writer, sheet_name="5_指标权重结果", index=False)        result_df_sorted = result_df.sort_values("排名").reset_index(drop=True)        result_df_sorted.to_excel(writer, sheet_name="6_综合评价排名", index=False)    print(f"✅ Excel文件已保存：{excel_path}")except PermissionError:    # 如果文件被打开，自动保存备份    backup_excel_path = os.path.join(SAVE_DIR, f"EWM-FCE_计算全过程_BACKUP_{timestamp}.xlsx")    # ... 同样写入操作    print(f"⚠️ 已自动保存到备用文件：{backup_excel_path}")

plt.figure(figsize=(12, 6))sorted_weight = weight_df.sort_values("熵权权重")colors = custom_pink_cmap(np.linspace(0, 1, len(sorted_weight)))plt.barh(sorted_weight["指标名称"], sorted_weight["熵权权重"], color=colors, edgecolor="white")plt.xlabel("熵权权重", fontweight="bold")plt.ylabel("指标名称", fontweight="bold")plt.title("各测量指标熵权权重排序", fontweight="bold")plt.grid(axis="x")plt.savefig(os.path.join(SAVE_DIR, "plots", "01_指标权重排序图.png"))

plt.figure(figsize=(10, 7))sorted_result = result_df_sortedcolors = custom_green_cmap(np.linspace(0, 1, len(sorted_result)))plt.barh(sorted_result["Treatment"][::-1], sorted_result["综合得分"][::-1], color=colors[::-1])plt.xlabel("模糊综合得分（越高越优）", fontweight="bold")plt.ylabel("处理组合（T×W）", fontweight="bold")plt.title("处理组合综合得分排名", fontweight="bold")plt.grid(axis="x")plt.savefig(os.path.join(SAVE_DIR, "plots", "02_组合得分排名图.png"))

plt.figure(figsize=(8, 6))heat_data = result_df.pivot_table(index="Factor1", columns="Factor2", values="综合得分", aggfunc="mean")sns.heatmap(heat_data, annot=True, cmap=custom_dual_cmap, fmt=".4f", linewidths=0.8, linecolor="white",            cbar_kws={"label": "模糊综合得分"})plt.title("T×W处理组合综合得分热力图", fontweight="bold")plt.xlabel("Factor2 (W)", fontweight="bold")plt.ylabel("Factor1 (T)", fontweight="bold")plt.savefig(os.path.join(SAVE_DIR, "plots", "03_TW交互热力图.png"))

plt.figure(figsize=(8, 8))top3_combinations = result_df_sorted.head(3)["Treatment"].tolist()plot_indicators = indicators[:8]   # 雷达图指标过多会拥挤，取前8个n_plot = len(plot_indicators)angles = np.linspace(0, 2 * np.pi, n_plot, endpoint=False).tolist()angles += angles[:1]ax = plt.subplot(111, polar=True)for i, treat in enumerate(top3_combinations):    treat_data = data_norm[df_mean["Treatment"] == treat][plot_indicators].values[0].tolist()    treat_data += treat_data[:1]    color = custom_pink_cmap(i / 3)    ax.plot(angles, treat_data, label=f"{treat}（排名{i+1}）", color=color, linewidth=2)    ax.fill(angles, treat_data, color=color, alpha=0.2)ax.set_xticks(angles[:-1])ax.set_xticklabels(plot_indicators, fontsize=9)plt.title("Top3处理组合指标得分雷达图", fontweight="bold", pad=20)plt.legend(loc="upper right", bbox_to_anchor=(1.2, 1.0))plt.savefig(os.path.join(SAVE_DIR, "plots", "04_Top3组合雷达图.png"))

print("="*60)print("✅ 熵权-模糊综合评价分析完成！")print(f"📂 所有文件保存路径：{SAVE_DIR}")print("="*60)best_result = result_df_sorted.iloc[0]print(f"🏆 最优处理组合：{best_result['Treatment']}")print(f"📊 综合得分：{best_result['综合得分']}")print(f"🔝 权重Top3指标：{weight_df.head(3)['指标名称'].tolist()}")print("="*60)print("🖼️ 生成图表列表：")print("  1. 01_指标权重排序图.png")print("  2. 02_组合得分排名图.png")print("  3. 03_TW交互热力图.png")print("  4. 04_Top3组合雷达图.png")print("="*60)

is_positive = {"成本": False, "缺陷率": False}data_pos, data_norm = data_preprocessing(df_mean, indicators, is_positive)