Indicate the number of months into the future for which you want to generate forecasts. This defaults to 12.

	The value chosen should be longer than the longest lead times for any stock code. This parameter can only be set at company level.

Indicate the default forecast calendar that must be used for forecasting (see Forecast Calendars).

You can override this entry either singly or for multiple stock codes at stock code/revision/release/warehouse level using the Options Maintenance facility in the Inventory Forecasting and Inventory Optimization modules.

When adjusting for outliers, the demand history for any sales period for a given stock code/revision/release/warehouse that differs from the mean by more than the figure entered here will have an adjustment entry generated. This defaults to 3.

	Periods with zero demand are not adjusted. This parameter can only be set at company level.

Any stock code/revision/release/warehouse that, when tested for seasonality, has a seasonal profile correlation of less than the figure entered here will be deemed to be non-seasonal. This parameter can range from 0 to 1.0 (where 0 corresponds to no correlation and 1.0 to complete correlation) and defaults to 0.75.

	This parameter can only be set at company level.

Select this to generate journals for amendments made to Families and Groupings structures using the Families and Groupings program.

This enables you to monitor information that has changed for structures such as Collection additions, changes and deletions and SKU (Stock Keeping Unit) additions and deletions. These changes affect the resultant forecast.

You can list the journals using the IO Amendment Journal program.

You can list the journals using the IO Selection Sets Amendment Journal program.

	This parameter can only be set at company level.

You can list the journals using the IO Policy Amendment Journal program.

	This parameter can only be set at company level.

You are prompted to perform the sales aggregation whenever you add to or delete an SKU, a consolidated warehouse item or a sub-collection in any collection, or when you maintain sales adjustments or adjust for outliers.

If you do not select this option, then you will not be automatically prompted to aggregate sales history in the Families and Groupings program. The Sales Aggregation screen will only be displayed when you select the Aggregate Sales History option from the Edit menu.

	Irrespective of your selection at this option, the IO Purge and Update program will perform the aggregation for you if an update is needed.

Select this to save the selections you made on all the tab pages and to exit the program.

Select this to print a report of your Inventory Optimization setup options.

It is advisable to keep this report for disaster recovery purposes.

Select this to exit the program without saving any changes you made.

Select this to view the latest online Help documentation for this program.

Indicate the default algorithm to use when forecasting. This defaults to Competition.

	You can override this at stock code/revision/release/warehouse level.

The following algorithms are available:

Competition uses all the algorithms in turn within the look back window (Number of months to use for competition defined below) to forecast what has already happened. It then compares the results of each algorithm with what the actual demand was and uses the selected statistical formula (Measure to use for competition) to determine which algorithm performs best. Choosing different values for either of the two parameters can result in a different algorithm 'winning' the competition.

Indicate the number of months into the past for which forecasts must be generated to compare algorithms. This defaults to 12.

	This parameter can be set as a default at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

Indicate the forecast error measurement that must be used when comparing algorithms. This defaults to Mean Square Deviation.

	This parameter can be set as a default at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

Select this to calculate the forecast error using the Cumulative Forecast Error (CFE) calculation.

This represents the difference between the forecast and the actual sale(s) for the period under review, displayed cumulatively.

Select this to calculate the forecast error using the Mean Absolute Deviation (MAD) calculation.

This is the average absolute deviation from the mean (i.e. the average error, ignoring the sign of the error).

Select this to calculate the forecast error using the Mean Square Deviation (MSD) calculation.

The sum of the squared forecast errors for each observation divided by the number of observations. It is an alternative to the Mean Absolute Deviation in that, because the errors are squared, more weight is placed on larger errors.

Select this to calculate the forecast error using the Mean Absolute Percentage Error (MAPE) calculation.

This is calculated as the average of the sum of all the absolute percentage errors for the data set. The absolute values of the percentages are summed and the average is computed.

That is, the difference between actual value and the forecast value is divided by the actual value. The absolute value of this calculation is summed for every forecast point in time and divided again by the number of fitted points. The result is displayed as a percentage.

Select this to calculate the forecast error using the Tracking Signal calculation.

This indicates whether the forecast average is keeping pace with any genuine upward or downward changes in demand.

The tracking signal is calculated as the sum of the forecast errors divided by the mean absolute deviation.

The mean (or average) is a very simple and robust algorithm. It is useful when:

This algorithm responds slowly to trends and does not detect seasonality.

Indicate the number of months (from 2 to 12) to include when calculating the mean. This defaults to 12.

If the calendar is defined in weeks, the calculation still uses the number of months defined here to calculate the equivalent number of weeks.

	This parameter can be set as a default at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

The median (or middle value) is a very simple and robust algorithm. It is useful when:

This algorithm responds slowly to trends and does not detect seasonality.

	If there are more periods with zero historic demand than with non-zero demand, the median will result in a forecast of zero, which may not be appropriate.

Indicate the number of months (from 2 to 12) to include when calculating the median. This defaults to 12.

If the calendar is defined in weeks, the calculation still uses the number of months defined here to calculate the equivalent number of weeks.

	This parameter can be set as a default at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

This is a simple algorithm that responds to trend. The smaller the value for the number of months, the more responsive the calculation will be to changes. It does not detect seasonality.

Indicate the number of months (from 2 to 6) to include when calculating the moving average. This defaults to 3.

If the calendar is defined in weeks, the calculation still uses the number of months defined here to calculate the equivalent number of weeks.

	This parameter can be set as a default at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

Also known as double exponential smoothing, this algorithm calculates the forecast by adding a data component to a trend component. Each of these components has its own coefficient. Depending on the value of the coefficient, the algorithm will respond more slowly or more rapidly to trends.

This algorithm does not detect seasonality.

Indicate the smoothing coefficient between 0 and 1.0) that must be used when calculating forecasts using this algorithm. This defaults to 0.7.

The coefficient determines the weighting applied to the last two historic demands. The most recent demand is multiplied by the coefficient while the older one is multiplied by one minus the coefficient.

	This parameter can be set as a default at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

Indicate the trend coefficient to use when calculating forecasts using this algorithm. This defaults to 0.6.

You can override this at stock code/revision/release/warehouse level.

This algorithm calculates the forecast by weighting the contribution of the 6 periods immediately prior to the current one. The periods will be weeks or months, depending on the calendar being used.

Depending on the weighting values, the algorithm will respond more slowly or more rapidly to trends.

This algorithm does not detect seasonality.

	The weightings for all periods added together must equal one. These parameters can be set as defaults at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

Indicate the weighting that should be applied to the two periods immediately prior to the period being forecast. This defaults to 0.5, meaning that these two periods contribute 50% to the calculation of the forecast.

Indicate the weighting that should be applied to the third and fourth periods prior to the period being forecast. This defaults to 0.3, meaning that these two periods contribute 30% to the calculation of the forecast.

Indicate the weighting that should be applied to the fifth and sixth periods prior to the period being forecast. This defaults to 0.2, meaning that these two periods contribute 20% to the calculation of the forecast.

This algorithm calculates the forecast by weighting the contribution of the 12 periods immediately prior to the current one. The periods will be weeks or months, depending on the calendar being used.

Depending on the weighting values, the algorithm will respond more slowly or more rapidly to trends.

This algorithm does not detect seasonality.

	The weightings for all periods added together must equal one. These parameters can be set as defaults at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

Indicate the weighting that should be applied to the first three periods immediately prior to the period being forecast. This defaults to 0.45, meaning that these three periods contribute 45% to the calculation of the forecast.

Indicate the weighting that should be applied to the fourth, fifth and sixth periods prior to the period being forecast. This defaults to 0.25, meaning that these three periods contribute 25% to the calculation of the forecast.

Indicate the weighting that should be applied to the seventh, eighth and ninth periods prior to the period being forecast. This defaults to 0.2, meaning that these three periods contribute 20% to the calculation of the forecast.

Indicate the weighting that should be applied to the tenth, eleventh and twelfth periods prior to the period being forecast. This defaults to 0.1, meaning that these three periods contribute 10% to the calculation of the forecast.

The two Holt-Winters algorithms (additive and multiplicative) are complex algorithms that use three smoothing equations for level, trend and seasonality.

	The coefficients for these three components of the forecast can be calculated by SYSPRO or set by the user.

In the Holt-Winters Additive algorithm, the forecast is calculated by adding the three terms together.

	These parameters can be set as defaults at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

Select this to let the system try to find the optimum coefficient values to use when forecasting, instead of the values at the coefficient fields below. This option is selected by default.

	Unless you have an advanced knowledge of how these algorithms work, we suggest that you let the system calculate the optimum values for the coefficients.

Indicate the level coefficient to be applied when forecasting with this algorithm. This defaults to 0.

Indicate the trend coefficient to be applied when forecasting with this algorithm.

Indicate the seasonal coefficient to be applied when forecasting with this algorithm.

The two Holt-Winters algorithms (additive and multiplicative) are complex algorithms that use three smoothing equations for level, trend and seasonality.

	The coefficients for these three components of the forecast can be calculated by SYSPRO or set by the user.

In the Holt-Winters Multiplicative algorithm, the forecast is calculated by multiplying the three terms together.

	These parameters can be set as defaults at company level and overridden for one or many stock code/revision/release/warehouse instances using the Forecast Calendars program.

Select this to let the system try to find the optimum coefficient values to use when forecasting, instead of the values at the coefficient fields below. This option is selected by default.

	Unless you have an advanced knowledge of how these algorithms work, we suggest that you let the system calculate the optimum values for the coefficients.

Indicate the level coefficient to be applied when forecasting with this algorithm. This defaults to 0.

Indicate the trend coefficient to be applied when forecasting with this algorithm.

Indicate the seasonal coefficient to be applied when forecasting with this algorithm.

Modeling allows the effect of a stock policy to be evaluated. The options below determine the outcome of the modeling calculation. By default, the model will take demand from the forecast. If no policy is in place, the suggested minimum will be zero and the suggested maximum equal to the total demand for the period.

The selections you make here are set as the defaults for the Modeling Options in the IO Stock Levels Modeling program and in the Information pane in the IO Forecast Accuracy Query program.

	These parameters can be set as defaults at company level and overridden for one or many stock code/revision/release/warehouse instances using the IO Stock Levels Modeling program. When changed within the IO Stock Levels Modeling program, the changes are treated as runtime values and are not saved.

The SYSPRO Inventory Optimization Suite uses a demand driven approach. This means that, in most cases, the inventory levels will be determined by the forecasted demand.

The parameter you select here is applied for each stock code/revision/release/warehouse instance (also called a SKU-Loc or Stock Keeping Unit Location).

This is the forecast used for the Requirements Calculation and is stored in the Requirements Planning tables within SYSPRO.

Sales order demand is used as the basis for the calculation to determine the minimum level required for each period.

For Make to Order items, ensure that the Demand Time Fences (DTFs) are accurately defined before selecting this option. Outside the DTF, the Highest of sales or forecasts, Sum of forecasts and sales, etc., rules are applied. Inside the DTF, only sales orders are considered. Therefore, if the DTF is zero, the rules are never applied. For a non Make to Order item, a Demand Time Fence cannot be defined, so a DTF at the beginning of the first period is assumed which means that the gross requirements rule is applied immediately. The DTF caters for the period it falls into even if its period length is shorter than the Inventory Optimization calendar period in use.

Select this rule to calculate the maximum stock level using the batching rule.

By default, with no policy in place the model will suggest a minimum of zero and a maximum equal to the total demand. If a policy is in place, the model will calculate the minimum and maximum according to the policy. In either case the calculation of the maximum will be modified by the effect of the batching rule.

	All of these parameters can be set as defaults at company level and overridden for one or many stock code/revision/release/warehouse instances using the Selection Set Maintenance program.

Please see Possible combinations of Comparison type and Forecast to use for the possible combinations of Comparison type with Forecast to use.

Select this to use the average of the most recent snapshots within the time window defined by the number of Snapshot months defined below.

	If the forecast calendar is in weeks, then the number of snapshots will be determined by how many weekly snapshots fall within the number of months defined in Snapshot months. If more than one forecast snapshot is created within a calendar period, only the last one is retained.

Select this to use the average of the most recent snapshots outside the lead time, within the time window defined by the number of Snapshot months defined below.

	If the forecast calendar is in weeks, then the number of snapshots will be determined by how many weekly snapshots fall within the number of months defined in Snapshot months. If more than one forecast snapshot is created within a calendar period, only the last one is retained.

Select the time window in months prior to the Run date selected for the Forecast Accuracy review. If the calendar used is in weeks, then all the periods that fall within this time window will be included. The time window can be anything from 1 to 12 months.

Select the time window in months for the periods used in the calculation of the moving average forecast in the Forecast Accuracy review. If the calendar used is in weeks, then all the periods that fall within this time window will be included. The time window can be anything from 1 to 12 months.

	This value is only used if the Comparison type selected is Moving average.

Indicate the time window, in months, prior to the Run date selected for the Forecast Accuracy review. If the calendar used is in weeks, then all the snapshots that fall within this time window will be included.

The time window can be anything from 1 to 6 months.

	This value is only used if the Forecast to use selected is either Average or Average outside lead.