Giters

daisiran / Final_Project

Monte Carlo Simulation Project

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Title: Random Inventory System Simulation

alt text

Team Member(s):

Xiaozheng Hou, Yinan Ni, Siran Dai

Monte Carlo Simulation Scenario & Purpose:

In the process of supply, it is always necessary to maintain a certain inventory reserve since the arrivals and sales cannot be the same amount and synchronized. If there are too many stocks, it will cause the funded backlog and the rise of custodial fees; if there are few stocks, it may lead to out-of-stock, resulting in loss of merchants' reputation and loss of customers. Therefore, we need to choose a suitable inventory and ordering strategy based on order policy.

We believe that using Monte Carlo Simulation will help us to find the lowest expected loss strategy and solve this issue.

So in our scenario, we will help the manager of a bicycles warehouse make the decision of choosing the best inventory strategy which yield the lowest cost.

alt text

In our first stage, we assumed that the warehouse have only one type of bicycle, and we compared and simulated several existing plans and picked the one with the lowest cost. After received feedbacks from the instructor and classmates, in the final stage, we decided to add the type of bicycles into our consideration as well as set the P(threshold of ordering) & Q(order amount) in a specific range based instead of some existing plans, which will make our scenario more realistic and applicable.

Simulation's variables of uncertainty

alt text

We decide the range the probability distribution by gaining data from relevant papers as well as doing primary research(one of our team member's relative is a bicycle warehouse manager, we had a interview with him). Although we believe there're more factors that need to be taken into consideration, we are confident that our current simulation will be helpful in the manager's ordering strategy decision-making process.

Hypothesis or hypotheses before running the simulation:

Hypothesis 1: If there is only one type of bicycle, the manager believe that there is no ordering strategy that will yield the cost less than 120000.

Hypothesis 2: If there are 5 types of bicycles, the manager believe that there is no ordering strategy that will yield the cost less than 600000.

Analytical Summary of your findings: (e.g. Did you adjust the scenario based on previous simulation outcomes? What are the management decisions one could make from your simulation's output, etc.)

Based on the results of our simulation, we found out that hypothesis 1 & 2 are wrong. There is at least 1 ordering strategy that will yield the cost less than the manager's expectation.

For only one type of bicycle, we also make a heatmap visualization that can clearly see different value of cost under the different combinations of P & Q. For more than one type of bicycles, we were unable to visualize the result. alt text

Instructions on how to use the program:

First, download and run the 'Final_version.py' file. Second, input iterate times and the number of types of bicycle.

All Sources Used:

About

Monte Carlo Simulation Project

Languages

Language:Python 100.0%