Exel
The below case study is written by Toby Krige of Exel Africa (formerly Tibbet & Britten Africa) and discusses how the company utilizes OPSI Systems' FLO Plan.
This article describes a fleet configuration exercise using both FLO and CAST. The study was done by Tobie Krige of Tibbett & Britten Africa (now Exel Africa), and we are grateful to him for sharing this article with us.
A. EXECUTIVE SUMMARY
1. Introduction
The objective of the exercise was to determine the core base-line fleet for the local delivery of the TBA Gauteng depot. The exercise was done on two modelling tools, namely CAST- dpm and FLO. The CAST system was used to indicate the number of trucks required for the local delivery of the TBA GAUTENG depot. The system has been used internally in the past and therefore was trusted as a good indicator. FLO was used to provide a more accurate answer with real life data.
2. CAST and FLO modelling software.
CAST-dpm
CAST-dpm is a strategic and tactical supply-chain modelling tool. It is designed to create a computer model of a complex distribution operation (dedicated or network based) and allows various strategies to be run against the model to find a least cost solution. CAST-dpm enables the examination of depot locations and the flow of product groups throughout the entire supply chain.
FLO
FLO is a vehicle scheduling and tracking system designed to improve fleet utilization and reduce delivery cost. FLO finds delivery routes and truckloads that minimize costs. It then compares the actual route driven to the suggested route using GPS satellite tracking. Any unauthorized vehicle use is automatically identified. By improving vehicle routing and reducing driver misuse of vehicles an overall reduction of up to 20% in delivery costs can typically be achieved. FLO also provides a fleet planning module called FLO-Plan.
3. Recommended Base Line Fleet
The recommended base line fleet was based on the low month volume throughput (August). The core fleet was a combination of various sizes of trucks (1, 4, 5, 8, 12, 24 & 34 ton). The operation people reached consensus and accepted the results of both simulation tools as workable and realistic. The results were further tested by an internal completion with the experts and many years of experienced employees to come up with the fleet size.
Although the two modelling software outputs were not the same it still provides an element of comparison and conclusion to be reached. The FLO plan showed an additional fleet saving of 15% compared with the CAST-dpm recommended fleet size. This is mainly due to more accurate scheduling and street level routing being done by FLO. The 3 largest vehicle types were almost the same in CAST and FLO. The 8 and 12-ton vehicles provided more diverse answers. Although the 12- ton vehicle was the most used vehicles in both modelling tools, CAST preferred more 1-ton vehicles than FLO, which selected the 4-ton vehicle for small deliveries. This was not a concern due to better and optimum routing in FLO.
The base fleet could be based on FLO results only if a routing and scheduling tool is being used for the planning and execution of the deliveries.
B. QUALIFICATIONS
- Data used. Data from the the last six months was used for this study. The data was manipulated to relate back to the base line throughput diagram for TBA GAUTENG depot. The data were consolidated by drop point per day. The data also provides quantities in units, mass and cubic meters per drop point. Other relevant information was addresses, store names and codes.
- NDD's. The Cast modelling was first done on actual data. This gave discrepant results with an over specification of 95%. This showed the inefficiencies for not managing NDD's (Nominated Delivery Days). Thereafter the NDD's were synchronised based on one drop per store per week for the FLO modelling tool. The results seem realistic and achievable.
- Operating parameters. The operating parameters were based on previous experience and calibration of similar product profile supply chains. The fixed standing time and variable standing could not be set by vehicle type in FLO and therefore time per drop and quantity were calculated and imported into FLO. Driver and crew operating hours were based on 12-hour shifts although delivery windows were set between 08h00 and 15h00.
- Constraints. Most of the constraints requested by the business were considered in the modelling. These included delivery windows, single drops and last drops.
- Vehicle capacities. TBA GAUTENG product profiles tend to be mass dominant. Therefore vehicle capacities were based on mass expressed in units.
- Coordinates. Longitude and latitudes were based on street level. Suburbs or town longitudes and latitudes were used when street addresses were not available.
- Seasonality. The throughput for the last six months in 2002 seems constant for the local deliveries, varying between 16% and 20% with August the lowest and October the highest.
- Cost. Although the first 4000 km per month were free, fuel cost was still being accounted for, to bring in a variable cost element in the selection of vehicle types.
- The base fleet. The outcome of the modelling process indicated that the relation between capacity, operating parameters and cost were very similar throughout the vehicle type selection process and therefore indicated that the selection of vehicle types was driven by drop sizes.
