Network design in supply chain management

Sue Abdinnour-Helm

International Journal of Agile Management Systems

1999, Vol. 1, No. 2, pp. 99-106

 

1)      Introduction

  1. a) 3 major components of supply chains
  2. i) Management of internal functions: processes used to transform raw materials to finished product

(1)   procurement

(2)   production

(3)   distribution

  1. ii) Management of upstream suppliers: ensures that the right material is received at the right time and to the right location

iii)     Management of downstream suppliers: ensures that customers receive the products they want in a timely manner

  1. b) Benefits of effective supply chain systems
  2. i) Reduced costs
  3. ii) Reduced inventory

iii)     Improved cycle time

  1. iv) Differentiation
  2. c) Proposition: Hub and Spoke network design
  3. i) Improve the distribution function

(1)   Lower cost

(2)   Better performance

  1. ii) Basic ideas

(1)   Consolidate traffic from different origins and sent it directly or via another hub

(2)   Products are insensitive to the number of hub stops they need to make

2)      Hub location

  1. a) Model Assumptions
  2. i) All hubs are connected
  3. ii) Each spoke is assigned to a single hub

iii)     All hubs have unlimited capacity

  1. b) Focus on the performance of two heuristics
  2. i) GA heuristic

(1)   Main idea: to employ the mechanics of natural selection to evolve initial solutions

(2)   A solution is represented by a string of 0s and 1s

(a)    Length of string=number of nodes in consideration

(b)   1s represent a hub, 0s spokes

(3)   Drawback: uses distance to assign spokes and hubs as opposed to traffic patterns

  1. ii) GATS heuristic

(1)   Introduced to overcome the shortcomings of using distance to assign hubs and spokes

(2)   Embeds a Tabu search (TS) within the genetic algorithm heuristic

(a)    TS – a search procedure that moves from one feasible solution to another

(b)   Exchange moves switch the assignment of a single city from one hub to another

(3)   This heuristic is proven superior in testing

3)      The optimal approach

  1. a) To minimize overall transportation cost subject to
  2. i) Number of hubs
  3. ii) Spokes are assigned to a single hub

iii)     A node becomes a hub if another node is assigned to it

4)      Data Sets

  1. a) 4 geographical layouts
  2. i) Continental USA
  3. ii) Regional USA

(1)   Eastern

(2)   Southwest

iii)     Canada

  1. iv) Western Europe
  2. b) Size of data sets is limited to 18 cities
  3. c) Distance between cities is the Euclidean distance between points as given by their coordinates

5)      Flow distribution

  1. a) Global Flow of goods
  2. i) Flow of goods is assumed to be inversely proportional to the distance between cities
  3. b) Localized flow of goods
  4. i) Flow of goods is assumed to be inversely proportional to the square of the relative distances between cities

6)      Conclusions

  1. a) Global flow has a lesser effect on the volume of flow between locations relative to local flow
  2. b) Models can help companies/logistics carriers to improve the design of their distribution function
  3. c) Results indicate that both GA and GATS give high quality solutions in both environments
  4. d) GA solution quality degrades more quickly than GATS in a local environment