The demand forecast solutions are based on the association of two powerful techniques: symbolic forecast and neural networks.
This association enables the forecast results to be sensitive to nonlinear effects in the time series as well as to incorporate a-priori expert or user knowledge.

This starting point enables the construction of forecasts solutions for demand in different areas ranging from the electric energy demand to cash demand at ATM.

AIA Group has developed the following taxonomy for its solutions, all of them based on the same forecasting technology.
Forecasting of Aggregated Demand
Forecasting of Distributed Demand

AIA Group also offers specialised forecasting consultancy, based on the considerable amount of expertise cumulated along its history.

Forecasting of Aggregated Demand.

The solutions for Aggregated Demand are characterised by:
· User with expertise in core business as well as in forecasting technology
· Excellent data quality
· High quality requirements on the forecast 
· Reduced data volume, office environment
· Number of series small (among 1 and 10)

AIA Group offers turnkey solutions for this kind of forecast, but points out the product for Aggregated Electric Demand Forecast, specialized in forecasting of hourly demand (active or reactive) for the aggregated load of a group of customers (a region or a zone).

The product for Aggregated Electric Demand Forecast offers the following characteristics for the planning and operation teams:
· Hourly demand of active and reactive power
· Reports designed for the analysis of daily, weekly and monthly demand.
· Real time connection (optional)
· Tools allowing the user to introduce and modify parameter values: hourly values, special days definition and temperature values.
· Analysis of historical series incorporating temperature, season and day of the week. Used for forecast and for the generation of the characteristics of the curves profile.

Installations at PG&E, ENDESA, REE, ONS

Forecasting of Distributed Demand.

The solutions for Distributed Demand forecasting are characterised by 
· Users with expertise in core business but, usually without deep knowledge of forecasting techniques.
· Uneven quality of data, often requiring a historical reconstruction.
· Forecast quality required, but balanced with costs.
· Large amounts of data. Client-server architecture or 3-tier (Web)
· High number of time series to analyse. (varying between 500 and 5.000)


AIA Group provides turnkey solutions for this kind of forecast.
To be noted:

· VIKING. A solution for the Gas Distributed Demand forecasting. The system performs forecasts for individual distribution networks, industrial customers and municipalities.
· APPOLO. Solution for Electric Distributed Demand forecast. The system performs individual and aggregated forecasts for High Voltage customers.

Systems installed at Gas Natural

The products MOKA and POLAR are Distributed Demand forecasting systems specialized in optimal cash management at bank branches and cash dispensers.