【SCM】基于报童模型的MAPE与Profit仿真

在销量预测中，提高预测准确率（1-MAPE）能否提高利润？

本文试图通过经典的报童模型进行随机仿真，观察MAPE和利润之间的关系。

一、模型设定

Newsvendor Problem -- 30 Days
Purchase cost	0.7
Selling price	1
Salvage value	0
Demand parameter μ	100
Amount ordered each day q	[80, 90, 100, 110, 120, 130]
Demand parameter ?σ	taking values from 1 to 30 with an increment of 0.1

二、仿真代码

import numpy as np
import matplotlib.pyplot as plt
import pandas as pd

#CONST
rounds = 30
c = 0.7   # purchase cost
s = 1     # selling price
u = 0     # salvage value
demand_mu = 100
color = ['b', 'g', 'r', 'c', 'm', 'y', 'k', 'w']
qs = [80]

# Demand
col = []
for demand_sigma in np.arange(1,30,0.1):
    row = [] 
    for _ in range(rounds):
        demand = max(round(np.random.normal(demand_mu, demand_sigma),0),0)
        while demand == 0:
            demand = max(round(np.random.normal(demand_mu, demand_sigma),0),0)
        row.append(demand)
    col.append(row)
col = np.array(col)
df_demand = pd.DataFrame(col, columns=[ i+1 for i in range(30)])

# newsvendor model 
fig = plt.figure(num=1, figsize=(10, 8),dpi=80) 
i = 1
for q in [80, 90, 100, 110, 120, 130]:
    df_sales_revenue = df_demand.applymap(lambda x: s * min(x, q))
    df_salvage_revenue = df_demand.applymap(lambda x: u * max(0, q-x))
    df_profit = df_sales_revenue + df_salvage_revenue - c * q
    df_mape = df_demand.applymap(lambda x: abs(x - q)/x)
    plt.subplot(2, 3, i)
    plt.scatter(df_profit.mean(axis=1), df_mape.mean(axis=1), c=color[i-1])
    plt.title('q = '+ str(q), size=26)
    plt.xlabel('Profit', size=12)
    plt.ylabel('MAPE',  size=12)
    i += 1

plt.legend(loc='best')

fig.tight_layout()#调整整体空白
plt.subplots_adjust(wspace =0.3, hspace =0.3)#调整子图间距
plt.savefig('./ppt素材/newsvendor_simulation.jpg') 

三、仿真结果图

结论：在各个订货点下， Profit随着MAPE的下降而升高，且趋于稳定。