Date of Publication

2005

Document Type

Master's Thesis

Degree Name

Master of Science in Industrial Engineering

Subject Categories

Industrial Engineering

College

Gokongwei College of Engineering

Department/Unit

Industrial Engineering

Thesis Adviser

Dennis E. Cruz

Abstract/Summary

Effective supply chain management in today's business world is considered a competitive advantage and supply chain managers today face a myriad of inter-related decisions ranging from inventory allocation to site selection to shipment scheduling. Unfortunately, most supply chain models consider such decisions independently, and therefore miss the effects of numerous tradeoffs. Managers thus face the lack of a tool by which they may evaluate the merits of novel integrative supply chain management paradigms such as lean logistics. The need for an integrative supply chain model for lean facilities was established. A mixed integer non-linear programming model was formulated for a supply chain with four echelons, each with multiple sites. The first echelon consisted of suppliers, then factories for the second echelon, down to depots and cross-docks in the third echelon, and finally, customer for the fourth echelon. The model made use of one-for-one base-stock replenishment policies for each facility and it also considered different modes of replenishment from suppliers to factories, namely, the traditional direct replenishment and the lean logistics mechanism of milk runs. The model also incorporated the choice between pull mechanisms Kanban and Constant Work-in-process or ConWIP. End-product demand generated by the customer echelon was modeled as stochastic. The objective was to minimize total system costs including capital and operating expenses, holding costs, transportation costs and backorder costs. The decision variables involved selection of sites for factories, depots and cross-docks, target inventory levels, replenishment frequencies and choice of pull system. In convexity analysis and model validation, the study focused on the normal probability distribution for the demand rates. Non-linear functions in the model were linearized via separable programming, while those that involved the integration of the normal density function were analyzed via Liebnizs Rule. The model was found to be convex but not strictly convex for z values greater than zero. Validation made use of the General Algebraic Modeling System or GAMS, particularly the DICOPT solver which alternated between solving the model as a nonlinear problem using CONOPT and as a mixed integer problem using CPLEX. The model was adjusted to consider the issues of scaling, variables bounds, non-linear sub-problem infeasibility, and initialization. In sensitivity analysis, designed experiments were used to ascertain relationships between the model parameters and the following responses: cost, milk runs, total system inventory, number of open facilities, and overall lean desirability. It was found, through Plackett-Burman screening designs, that transportation cost, holding cost and demand variability were the parameters that most significantly affected the optimal solution. Central composite designs for Response Surface Methodology established the configurations of these parameters where lean mechanisms would be desirable. Overall, the desirability of lean systems was found to be heavily dependent on transportation cost. For low transportation cost, lean systems are desirable in low demand variability, high holding cost environments. If transportation costs are high, then holding costs must be low in order for lean systems to be most desirable. Suggested directions for further research include the consideration of supply-side variability, as well as direct transportation from factories to customers, foregoing the echelon for depots and cross-docks.

Abstract Format

html

Language

English

Format

Electronic

Accession Number

CDTG003887

Shelf Location

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

Physical Description

1 computer optical disc ; 4 3/4 in.

Keywords

Inventory control; Business logistics; Facility management; Marketing--Management

Upload Full Text

wf_yes

Share

COinS