Model Optimisation Flashcards
What is Optimisation?
How do existing stores compare with the optimal stores? i.e. a store in every location
What is the issue with optimisation?
Combinational Problem - e.g. want 5 new stores in 100 zones gives 50 mil. solutions - too large for SIMs to compute
What are the two scenarios for optimisation?
Ideal Representational Plan (IRP)
Mergers and Acquisitions
How does the IRP work?
- Create 100 zones and try a store in every zone for max. revenue. Example answer = zone 56
- Fix location in 56 and try other 99 locations for next best. E.g. answer = 27
- Fix locations 56 and 27 - repeat until have 5 stores
What is the difference between a Full IRP and a Rationalised IRP?
Full: maintains all existing no. stores but relocates them to the ideal locations to increase sales and max. revenue
Rationalised: reduces the no. stores to reduce costs whilst maintaining or increasing sales - the ideal locations with the fewest sites
What is the ideal IRP?
RATIONALISED - gives the ideal blueprint for growth
However, as more companies get hold of the tech. they can all create their ideal blueprint
What are the 3 optimisation methods for mergers?
- Pre-merger planning - who’s the optimal partner?
- Check competition commission - by acquiring/merging, what’s the impact on local competition?
- Post-merger planning - optimising new/combined networks
What went wrong with Barclay’s acquisition of Woolwich in 2000?
Acquired for £5.3bn
Barclay’s plan: close any Woolwich store within 100m of existing Barclay’s branches - 100 closures of 400 stores
Sold the 100 off
BUT didn’t research success of Woolwich stores, therefore ended up shutting some of the most successful ones
What did Clarke et al. (2003) create in response to the Barclay’s failure?
8 scenario optimisation model for the financial services sector in order to give each town a geographically unique scenario
How can GIS be used to find optimal locations?
Location-Allocation Models by finding the optimal supply point and then trying to minimise the euclidean distance - but this method is limited, therefore argue that GIS is not best suited for optimal location (BIRKIN ET AL., 1999)