Show percentage of ETRI EBID expected price for best beading using monte carlo simulation.

etri_ebid_expected_price.py

# https://ebid.etri.re.kr/ebid/
# 6. 예비가격 및 예정가격
#   ① 복수예비가격은 전자입찰 시작 직전에 전자입찰시스템의 난수발생기를 통하여 기초금액(예비가격
#        기초금액)의 ±2% 범위 내에서 15개의 예비가격으로 생성됩니다.
#   ② 예정가격은 전자입찰자가 응찰조서작성시 추첨한 예비가격번호 중 가장 많이 추첨된 번호순으로 4개를
#        선정하여 각 번호에 해당하는 예비가격을 산술평균하여 결정됩니다
#
# 나라장터 복수예비가격 선정 방법 (더 자세히)
# http://www.g2b.go.kr/pt/e-support/fwdEsupportMain.do?supportUrl=/pt/notice/listNotice.do 에서 예비가격 검색
# 
# 예시: 100000번 시뮬레이션, 2명의 입찰자, 예비가격: 51,100,000원
# 결과 화면: https://i.imgur.com/rnhURZr.png

import matplotlib.pyplot as plt
import pandas as pd
import random
import sys
from collections import Counter

def generate_price_list(price, percent, count):
    # generate reserved price list between 98% ~ 102%    
    # p = [random.uniform(price*(1.0 - percent), price*(1.0 + percent)) for i in range(count)]
    # generate reserved price list base on table from g2b.go.kr (나라장터)
    # 복수 예가 작성 기준 (reserved price list table) https://i.imgur.com/KbWM0T5.png
    p = [random.uniform(price*((1.0 - percent) + (percent*2/count*i)), price*((1.0 - percent) + (percent*2/count*(i+1)))) for i in range(count)]
    return p

def ebid_sim(num_simulation, bid_participants, reserved_price):
    final = []
    for j in range(num_simulation):
        price_list = generate_price_list(reserved_price, 0.02, 15) # ±2%, 15 reserved prices
        prices = []
        for i in range(bid_participants):
            for p in random.sample(price_list, 2): # choose two from price_list
                prices.append(p)
        prices_dict = dict(Counter(prices))
        sorted_prices = [v for (v, c) in sorted(prices_dict.items(), reverse=True, key=lambda item: item[1])[:4]] # best 4 in most order
        average_price = sum(sorted_prices)/len(sorted_prices) # average of four prices.
        final.append(average_price)
    return final

if __name__ == "__main__":
    reserved_price = 51100000.0
    if len(sys.argv) > 1:
        reserved_price = float(sys.argv[1])
    else:
        print("사용법: python3 etri_ebid_expected_price.py [예비가격]")
        print("예시: python3 etri_ebid_expected_price.py 51100000")
        print(f"현재 기본 예비가격: {reserved_price:,.0f}원으로 시뮬레이션을 실행합니다.\n")

    plt.style.use('ggplot')
    final = ebid_sim(100000, 2, reserved_price) # 100000 loop, 2 bid participants,  51,100,000 reserved price 
    output = pd.DataFrame(final)
    ax = output.plot(kind='hist', bins=30)
    ax.set_xlabel("Price")
    ax.set_ylabel("Frequency")
    print(output.describe(percentiles=[p/20 for p in range(1, 20)]))
    plt.show()

count 1.000000e+05
mean 5.109848e+07
std 2.769738e+05
min 5.015367e+07
5% 5.063944e+07
10% 5.073822e+07
15% 5.080506e+07
20% 5.085927e+07
25% 5.090678e+07
30% 5.094902e+07
35% 5.098880e+07
40% 5.102591e+07
45% 5.106299e+07
50% 5.109903e+07
55% 5.113449e+07
60% 5.117102e+07
65% 5.120848e+07
70% 5.124765e+07
75% 5.129035e+07
80% 5.133759e+07
85% 5.139175e+07
90% 5.145852e+07
95% 5.155631e+07
max 5.200070e+07