planfocus has built forecasting and planning algorithms with a focussed team of experts from the areas of forecasting, mathematical optimization, operations research and computer science, over the course of over a decade. We have custom-built an arsenal of highly specialized optimization methods, some patent protected, which CCO® draws from, depending on the specific scenario.
Cash Point Forecasting
Forecasting is an advanced, adaptive, self-learning statistical method that is based on the cash demand per denomination, drawing both from long-term as well as intraday data for each cash point and denomination. It uses observed seasonal activity patterns, monthly payday patterns, configured event-days (configured within regional or cash point group calendars). If historical data is present, no manual input is required for event-activities, as this is fully automatically derived from the historic information. Forecasting is treating cash point data of withdrawals and deposits separately with any cash recycler and branch device, so that independent flows can be optimally forecasted.
Cash planning fully automatically generates cash orders that are presented for final confirmation. Cash orders are usually generated during the night in order to optimally use compute recourses and in order for cash managers to be able to immediately work on cash orders in the morning hours. Usually none or very few modifications are necessary, which is why for CCO® we use the term “order” instead of “advice”, although each order can be manually modified.
Planning of cash replenishment actions happens down to the hour, days and weeks in advance, dynamically considering possible service days, replenishment times, ATM equipment behavior, substitution effects, and much more, on the basis of the forecasts, equipment configuration and business constraints, using patented discrete mathematical optimization algorithms from field of artificial intelligence.
Cash planning uses a multitude of parameters that are “calibrated” (fine-tuned) during the rollout configuration to the specific client situation at hand. Our business analytics team helps you to initially configure the calibration during the initial Go-Live process. Over the course of larger implementations, you will be given full control to continue the process and maintain this parameter configuration very efficiently over time.
Deep Optimization of Groups
In addition to the optimization of individual cash points, CCO® employs numerous specialized algorithms to optimize entire groups of cash points together – however, always at the same detailed level of the individual denomination and considering all intra-day effects. Examples include: Synchronizing multi-cash-point locations (branches and remote locations, module CCO/Syncro) using dedicated synchronization methods. Balancing of optimal cash-point tour allocations (module CCO/Transport).
During the daytime, cash point inventories are permanently observed (CCO/LiveInventories) and CCO® can generate alert conditions and emergency orders, should an unexpected cash shortage be upcoming. Over time, costs and quality of the overall process can be monitored by the management using the visual CCO® Management-Cockpit, which allows for charting important statistics and trends, such as:
- Cash orders acceptance ratio (i.e. manual changes)
- Cash orders not executed as ordered (i.e. CIT timing/alterations)
- Remaining inventories (“Rebanks”) over time
In addition, each cash order can be inspected in terms of its remaining cash upon execution as opposed to the forecasted remaining cash.