A supply chain model for lean facilities considering milk run replenishment, postponement production system and one-for-one base stock inventory policy decisions

Date of Publication

2009

Document Type

Bachelor's Thesis

Degree Name

Bachelor of Science in Industrial Engineering

Subject Categories

Engineering

College

Gokongwei College of Engineering

Department/Unit

Industrial and Systems Engineering

Thesis Adviser

Dennis E. Cruz

Defense Panel Member

Ronaldo V. Polancos

Abstract/Summary

In today's competitive world and the globalization of the market economy, competition has moved towards a macroscopic competition that is between supply chains from the past microscopic competition against other firms. In addition to this increasing competitiveness, customer's preference shifted to more personalized products to suit their everyday needs and wants. Thus effective supply chain management is increasingly becoming an effective tool to gain competitive advantage. Unfortunately most supply chain models considered only a combination of strategic and strategic operational decisions in their supply chain and foregoing the tradeoffs of considering only the operational decisions such as flexibility and optimization of daily activities that would minimize cost. Furthermore, in the growing field of supply chain management, the lean and agile paradigms are becoming increasingly popular. Yet again most supply chain models considered only either lean or agile manufacturing separately thus missing numerous tradeoffs. Thus the need for a supply chain model that considers only operational decisions and the combination of the lean and agile paradigms through postponement is established.

A mixed integer non-linear programming model was formulated for a supply chain with four echelons, each with multiple sites and considering multiple products. The first echelon consists of the suppliers of raw materials, manufacturing plants for the second echelon, warehouses and cross-docks for the third echelon and the end customer for the fourth echelon. Bayesian forecasting method was used to forecast the demand which will aid in the decision of postponement alongside the perceived customer demand uncertainty. Kanban pull systems are installed in the manufacturing plants while ConWIP are installed for the warehouses. The model also considers cross-docking and milk runs which are consistent with lean. The decision variables are the time between replenishments and the mode of replenishment of the manufacturing plants, the proportion of demand, the base stock levels at the manufacturing plants and the warehouses, the milk run set, and the proportion of the demand to be postponed. The model was validated through the GAMS software, particularly the DICOPT solver which alternated between solving the model as a non-linear problem using CONOPT and as a mixed integer problem using CPLEX. A small model was created, and variables were scaled and bonded to help in the validation.

In the sensitivity analysis, designed experiments were used to analyze relationships between the parameters and the following responses. It was found through the two level fractional factorial screening design that the raw material and intermediate product holding cost, the milk run variable cost, the mean demands per customer per product, the actual due dates for each customer for each product and the service levels are the parameters that are most significant. Central composite designs for Response Surface Methodology established the configurations of the parameters that answered the sub-problems of the study. The desirability of postponement was affected by the demand quantity and demand variability. The relationship is inversely proportional for demand quantity, the lower the demand the more items would be postponed and vice versa. For demand variability, the relationship with postponement is directly proportional. Higher variability of demand would lead to a higher proportion of finished goods to be postponed and vice versa. Total system cost will be minimized when the demand is lowest as well as the raw material and intermediate product holding costs. Furthermore in terms of direct fixed costs, total system costs increases as direct fixed costs increases. Milk run desirability is affected by the actual due dates of the customers and the RM holding costs and direct fixed costs. Milk runs are desirable when actual due dates are low and high. Furthermore, milk runs would be more desirable when RM holding costs and Direct replenishment fixed costs are high. Total system inventory is affected by the demand, RM and FG holding costs and milk run variable costs and direct replenishment fixed costs. In terms of RM and FG holding costs, the relationship with total system inventory is inversely proportional. In terms of demand, more demand means more inventory to satisfy customer orders. Furthermore, having higher milk run variable costs than direct replenishment fixed cost increases the total inventory of the syste

Recommendations for further study include capacity and stochastic lead times for suppliers. For manufacturing plants, single production line, and set-up time to be considered for multiple products that aren't part of a single family. For replenishment mode of manufacturing plants particularly milk runs, vehicle routing could be considered to further minimize the cost. For the warehouse replenishment, direct replenishment and milk runs could be considered and finally the study can be extended to multiple periods to further see the effects of forecasting.

Abstract Format

html

Language

English

Format

Print

Accession Number

TU15331

Shelf Location

Archives, The Learning Commons, 12F, Henry Sy Sr. Hall

Physical Description

xi, 298 leaves : col. ill. ; 28 cm.

Keywords

Business logistics--Cost effectiveness; Supply and demand; Production management; Delivery of goods--Management

This document is currently not available here.

Share

COinS