Demand Forecasting in Supply Chain Management

Description

TXT e-Solutions currently uses TXTDemand for demand forecasting in its Supply Chain Management solution.

TXTDemand combines sophisticated forecasting algorithms as well as algorithms for the analysis of historic data and current sales data with interactive revision tools.

Current drawbacks:

• "batch": huge amounts of data have to be analysed and digested daily by complex algorithms (up to 100GB) mostly keeping expensive IT resources busy for about 8 hours every night;
• mostly implemented by centralised clusters of expensive systems (1,5 M Euro)
• Computational process’s duration takes about 1 day (data gather, run batch)
• slowly changing computational power request
• "interactive" phase: humans locally adopt collaborative decision support dashboards (business intelligence) to over-write some forecasts, to propagate them and to perform what-if analysis and simulations
• strict performance requirements: response times for the most frequent types of interaction due to reduced computational power of client.
• Fast moving computational power request

Desired improved architecture & performances using SORMA:

• parallel/high-performance computation
• distributed data-storage supporting parallel computation
• more scalable architecture to meet different customer’s environments (budget) based on Globus Toolkit 4

Integration scenario with SORMA:

• The BATCH phase implies "slowly" changing request of GRID resources, since the number of records to process is more or less stable over time.
• there can be small variations daily or weekly, requiring the negotiation of new resources on the GRID market
• During the INTERACTIVE phase, the request for GRID resources depends on daily human activities,
• there can be peaks of requests, happening in a short time.
• This is more challenging for SORMA
• In this case, negotiations/bidding on the GRID market & dynamic allocation of resources should happen in nearly real-time

News