|
Case study - Petroleum Industry Distribution Optimization
System
Bob Blightman - Cirrus
System Overview
The system was developed by Cirrus, building on its extensive experience
of modeling all aspects of distribution within the food, pharmaceutical
and retail sectors. The system is generally customized to meet each
user's special requirements although there is a central optimization
core around which each installation is based. Usually the customization
process takes only a short time and the system can be installed
within a 2-4 week period, including training and documentation.
The system is now in use at some 25 user sites in the UK, the rest
of Europe, Brazil, South Africa and India where it is regularly
used to ensure the "optimum" supply chain strategy is being adopted.
The objectives of the primary and secondary distribution optimization
system are as follows:-
- To identify the optimum (least cost) way
of delivering forecast demand to retail and commercial sites
- To calculate the optimum (least cost) way
of delivering demand to the depots from the primary source through
pipeline, sea, rail or road
- To allocate delivery points to the least
cost depot
- To analyze effect of closing a depot
- To analyze sensitivity of costs to marketing
activity
- To evaluate exchange agreements
- To calculate number of vehicles and drivers
required at a depot
- To identify potential capacity constraints
at depots
- To calculate total cost of delivering product
- To calculate the cost of a constraint
In calculating the "optimum" solution, the model takes
into account capacity constraints at specific depots, product availability,
vehicle bases and back hauling opportunities. Optimization is achieved
through the use of Xpress-MP, seamlessly
embedded within the total system. The model is able to deal with
"step function" costs (e.g. as volumes increase, costs per litre
may decrease), and 0/1 variables (e.g. should we close any depots?)
The model is driven by a number of sets of data that
describe the location of customer sites, sources of supply, depots
and relevant physical constraints. The main elements of input data
are:
Product data
Sources of Supply data, including refineries, exchange
agreements and purchases, reflecting capacities, cost relationships
and the available linkages to the depots
- Depot Data, reflecting capacities and cost
relationships
- Customer Data
- Demand Data
- Time and Distance Matrix, which can be generated
through the use of other tools embedded within the system e.g.
Mapinfo.
The latest version of the system includes an ability
to model multi-drop operations. The system provides a wide range
of results following the optimization process. These results are
available as either numeric values or in the form of graphical reports.
The system calculates the total cost of the solution and lists the
depot that each customer has been allocated to.
The system makes use of a mapping module that displays
a map of the relevant country and pinpoints the location of each
depot and delivery point. The user can "click" onto any one of the
points for information about the selected depot of delivery point.
The user can also select and zoom into specific areas
of the country for a more detailed representation of an area. The
delivery points are colour coded to indicate which depot they have
been allocated to. The user can also select specific display criteria
to highlight sites possessing particular criteria. For example it
is possible to display all sites that cost more than a specified
cost per litre to deliver, or show the most profitable sites. Other
reports show how much volume has been passed through each depot
shown as a % of capacity.
Alternatively, the system can provide output such
that it can be imported into other GIS tools (e.g. MapInfo)
- Other reports include:
- Supply cost to each customer per litre
- Delivery cost
- Trip Distance and time by customer
- Customers share of fixed costs
- Customers total annual cost
- Comparison with previous allocation
- Total tonnage by depot by grade plotted against
capacity
- Total trips per week and vehicles required
- Maximum distance traveled from depot
- Number of drivers/vehicles required
- Amount supplied by each source
- The value of exchange agreements
- Sensitivities on constraints
For more information, please contact Bob Blightman
at bb@cirrusconsulting.com, telephone (+44) 1252 795610
|
