What is the difference (if any) between SCM and Business Logistics Management?
Supply Chain Definition (G.C. Stevens, 1989): “. . . a connected series of activities which is concerned with planning, coordinating and controlling materials, parts, and finished goods from supplier to customer. It is concerned with two distinct flows (material and information) through the organization .”
The Basic Problem: Get the right amounts of the right products to the right markets at the right time in the most economical way.
We can’t forget additional factors related to product design for manufacture and distribution, i.e., the constraints that product characteristics place on ease of manufacture and distribution. Another important issue is product mix from a marketing standpoint, i.e., which products the chain will carry.
Consolidation – if the market demands several products made by the manufacturer, consolidating them into one warehouse will make it more economical to send frequent consolidated shipments of full truckloads to the market.
Mass Customization – a modular product architecture helps enable mass customization, which is the ability to mass produce goods that can quickly and easily be customized to individual specifications (as in the Gateway and Dell computer examples).
JIT/VMI - Just-in-time and vendor managed inventory strategies to smooth flow of goods and increase response time of suppliers.
Order cycle time: time between placing order and receiving product
Order processing and assembly
Additional stock acquisition time (if out of stock)
“ On the average it is approximately six times more expensive to develop a new customer than it is to keep a current customer. Thus, from a financial point of view, resources invested in customer service activities provide a substantially higher return than resources invested in promotion and other customer development activities.”
P.S. Bender, Design and Operation of Customer Service Systems , 1976.
Cost vs. Service Models Customer Service Level Revenue (sales) Profit Logistics costs
Transport involves equipment (trucks, planes, trains, boats, pipeline), people (drivers, loaders & unloaders), and decisions (routing, timing, quantities, equipment size, transport mode). In underdeveloped countries we often find it necessary to locate production close to both markets and resources, while in countries with developed distribution systems people can live in places far from production and resources.
When deciding the transport mode for a given product there are several things to consider:
Once we have selected a transport mode and have goods that need to go from point A to point B, we must decide how to route a vehicle (or vehicles) from point A to point B.
Given a map of all of our route choices between A and B we can create a network representing these choices The problem then reduces to the problem of finding the shortest path in the network from point A to B.
This is a well solved problem that can use Dijkstra’s Algorithm for quick solution of small to medium (several thousand nodes) sized problems.
Suppose we have multiple sources and multiple destinations, that each destination requires some integer number of truckloads, and that none of the sources have capacity restrictions. In this case we can simply apply the transportation method of linear programming to determine the assignment of sources to destinations.
A second heuristic, known as the sweep heuristic, will perform much better in the ‘worst case’ then the nearest neighbor.
The sweep heuristic basically attempts to make an outer loop around the nodes.
Draw a straight line emanating from the depot with a length r which is at least as great in length as the maximum straight-line distance from the depot to any customer (the direction of the line is not important).
Visualize the line as sweeping either clockwise or counter-clockwise through a circle of radius r . Each time the radius line intersects a customer location we make that customer the next customer on the route.
Single Depot, Multiple Destinations, Vehicle Capacities
When the depot contains many vehicles and vehicle capacity constraints come into play, the problem becomes even more complex.
If each customer has enough demand to receive a full truckload the problem is easy and we simply use the shortest path to get the single truck to each customer. Otherwise, we must decide which customers will receive deliveries from the same truck, and then we must decide how to route the trucks to the customers on the route.
We will look at a mixed-integer programming formulation of the Vehicle Routing Problem (VRP).
Illustration of VRP (Outlier) Depot 50 76 39 112 88 29 123 44 58 90 77 89 57 115 124 59 176 65 98 125 Truck Capacity = 250 What is the minimum # of trucks we would need? Maximum?
The Vehicle routing problem (VRP) generalizes the TSP since we have a set of K capacity constrained (homogeneous) vehicles at a depot, each of which must visit a subset of the n - 1 customers exactly once and return to the depot.
No two vehicles may visit the same customer. This means that each vehicle must complete a Hamiltonian tour (a Hamiltonian tour is a feasible TSP solution).
The objective is to determine the minimum travel cost required to serve each customer. Let A denote the set of pairs of cities, and let k index trucks, each with capacity u . Assume that customer i has demand equal to d i .
Given the difficulties in solving the TSP problem, we cannot expect to have great success solving VRP problems without some sort of heuristic approaches. We can use several guiding principles in developing these heuristics. (Note that the above formulation does not consider additional practical restrictions such as limits on driver time, time window delivery restrictions, or return of goods from customers to the depot.)
Note that the sweep method, when applied to the VRP, will have a slightly different interpretation. That is, we can only add a delivery location to a route as long as it does not exceed the vehicle capacity. So we can only continue to assign deliveries to a route as long as the vehicle capacity is not exceeded. Then we need to start assigning deliveries to a new truck.