Hybrid NSGA-II Algorithm on Multi-objective Inventory Management Problem - Intelligent Computing for Sustainable Energy and Environment - page 160

Hybrid NSGA-II Algorithm on Multi-objective
Inventory Management Problem
Lin Lin and Shiji Song
Dept. of Automation, TNList, Tsinghua University, Beijing 100084, P.R. China
[e-mail], [e-mail]
Abstract. Inventory management is a key issue in supply chain man-
agement. Under the circumstances that there are plenty of risks, it is
more usable and appropriate if the risk problem is also taken into con-
sideration when addressing the issue of inventory management. In this
paper, we firstly introduces the classifications of inventory model, in-
troduces two parameters, VaR and CVaR to measure risks. Also, we
established a bi-objective model considering inventory cost and CVaR at
the same time. Heuristic method to solve the problem is addressed then.
We examined the application of Genetic Algorithm on multi-objective
problems, i.e. the NSGA-II algorithm. We proposed an analytic method
to simplify the solution of the problem. Besides, we examined the local
search method based on the problem and proposed a Hybrid Genetic Al-
gorithm. Simulation verifies the usability of our model and the eciency
of our algorithm.
1 Introduction
Supply chain is a network that integrates the suppliers, manufacturers, retails
and the end customers. And to manage a supply chain is to manage the flow
of raw materials, products and information flow that involves. It is shown by
research that inventory system occupies 20% − 60% value of all the manufacturing
industry, making inventory management a focus in supply chain management.
Inventory management is to control the behavior of when to order and how much
to order so that the manufacturing process can be sustained and products can
meet the needs of retailers or customers.
As to the supply chain that consists of many layers, there is much risk to
be considered. Risk can be from exterior or interior. The delay in supply and
less demand from customers are both interior risks [1]. Terrorism and SARS
are on the other hand the exterior ones. These risks exert great influence on
the company or the entire supply chain. Therefore, it is meaningful to add risk
dimension to inventory problem. This passage in organized like this. Section 2
introduces the literature about inventory models and risk measurements. Section
3 builds a bi-objective model which can show the inventory cost and risk at the
same time. At last, section 4 proposes a problem-based hybrid algorithm to solve
this problem. Through simulation, we verified the feasibility and eciency of this
algorithm.