MRP Statistical Forecast Shows Unexpected Results/Forecast Values

(Doc ID 2337943.1)

Last updated on DECEMBER 11, 2017

Applies to:

Oracle Materials Requirement Planning - Version 12.0.4 and later
Information in this document applies to any platform.

Symptoms

On : 12.0.4 version, Forecast in Test:

Users are testing an Inventory Statistical Forecast. They notice in the output they see the same value for forecast for each period after running the Generate Forecast - INCFIF concurrent request. They expected to see different values for the period buckets. Example data:

Forecast Rule setup
Name = P - Statistical
Bucket Type = Periods
Include = Miscellaneous Issues
Forecast = Statistical
Maximum Past Periods = 24
Alpha Smoothing Factor = 0.1
Seasonality index = 1 for all 13 periods

Forecast Set name = USC_Periods
Forecast = Top-Items

Generate forecast parameters
Forecast Name = Top-Items
Forecast Rule = P - Statistical
Selection = Specific Inventory Item
item = 781608002
Overwrite = No
Start = 15-NOV-2017
Cutoff Date = 15-May - 2019

The log file with profile MRP:Debug Mode = Yes shows the following data:

The value of profile option 'MRP_ROUND_SOURCE_ENTRIES' is N
The value of profile option 'MRP_IGNORE_ZERO_QTY_DEMANDS' is Y
SELECT MAX(PERIOD_START_DATE) FROM BOM_PERIOD_START_DATES D, MTL_PARAMETERS P WHERE ORGANIZATION_ID=101 AND P.CALENDAR_CODE = D.CALENDAR_CODE AND D.EXCEPTION_SET_ID = P.CALENDAR_EXCEPTION_SET_ID AND D.PERIOD_START_DATE <= SYSDATE ORDER BY 1 DESCSELECT TO_NUMBER(TO_CHAR(D.PERIOD_START_DATE,'J')),D.PERIOD_SEQUENCE_NUM FROM BOM_PERIOD_START_DATES D, MTL_PARAMETERS P WHERE P.ORGANIZATION_ID=101 AND P.CALENDAR_CODE = D.CALENDAR_CODE AND D.EXCEPTION_SET_ID = P.CALENDAR_EXCEPTION_SET_ID AND (D.PERIOD_START_DATE >='30-OCT-17' AND D.PERIOD_START_DATE<= fnd_date.canonical_to_date('2019/05/15 00:00:00')) ORDER BY 1
Input to BuildDmds, cur_cal_date: 02-OCT-17
SELECT NVL((0+MISCELLANEOUS_ISSUE),0), TO_NUMBER(TO_CHAR(D.PERIOD_START_DATE,'J')),D.PERIOD_SEQUENCE_NUM FROM MTL_DEMAND_HISTORIES H, BOM_PERIOD_START_DATES D, MTL_PARAMETERS P WHERE P.ORGANIZATION_ID=:org_id AND P.CALENDAR_CODE = D.CALENDAR_CODE AND D.EXCEPTION_SET_ID = P.CALENDAR_EXCEPTION_SET_ID AND H.PERIOD_START_DATE(+)=D.PERIOD_START_DATE AND H.ORGANIZATION_ID(+)=:org_id AND H.PERIOD_TYPE(+)=3 AND H.INVENTORY_ITEM_ID(+)=:item_id AND (D.PERIOD_START_DATE <= '02-OCT-17' AND D.PERIOD_START_DATE >= (SELECT NVL(MIN(PERIOD_START_DATE),TO_DATE('01-01-1900', 'DD-MM-YYYY')) FROM MTL_DEMAND_HISTORIES WHERE ORGANIZATION_ID=:org_id AND PERIOD_TYPE=3 AND INVENTORY_ITEM_ID=:item_id)) ORDER BY 2 DESCStatistical Forecast in TSEFModel function
  forecast, dmd[i-1], base0, base1, *trend, season[j-1], prd_seq_num
forecast = 305.000000 305.000000 305.000000 305.000000 0.000000 1.000000 12
forecast = 297.000000 225.000000 305.000000 297.000000 0.000000 1.000000 1
forecast = 290.400000 231.000000 297.000000 290.400000 0.000000 1.000000 2
forecast = 297.160000 358.000000 290.400000 297.160000 0.000000 1.000000 3
forecast = 290.544000 231.000000 297.160000 290.544000 0.000000 1.000000 4
forecast = 280.589600 191.000000 290.544000 280.589600 0.000000 1.000000 5
forecast = 277.830640 253.000000 280.589600 277.830640 0.000000 1.000000 6
forecast = 271.147576 211.000000 277.830640 271.147576 0.000000 1.000000 7
forecast = 266.732818 227.000000 271.147576 266.732818 0.000000 1.000000 8
forecast = 273.259537 332.000000 266.732818 273.259537 0.000000 1.000000 9
forecast = 268.833583 229.000000 273.259537 268.833583 0.000000 1.000000 10
forecast = 258.650225 167.000000 268.833583 258.650225 0.000000 1.000000 11
forecast = 257.385202 246.000000 258.650225 257.385202 0.000000 1.000000 12
forecast = 249.546682 179.000000 257.385202 249.546682 0.000000 1.000000 1
forecast = 242.392014 178.000000 249.546682 242.392014 0.000000 1.000000 2
forecast = 249.152812 310.000000 242.392014 249.152812 0.000000 1.000000 3
forecast = 255.437531 312.000000 249.152812 255.437531 0.000000 1.000000 4
forecast = 252.493778 226.000000 255.437531 252.493778 0.000000 1.000000 5
forecast = 242.844400 156.000000 252.493778 242.844400 0.000000 1.000000 6
forecast = 233.759960 152.000000 242.844400 233.759960 0.000000 1.000000 7
forecast = 227.283964 169.000000 233.759960 227.283964 0.000000 1.000000 8
forecast = 235.055568 305.000000 227.283964 235.055568 0.000000 1.000000 9
forecast = 228.250011 167.000000 235.055568 228.250011 0.000000 1.000000 10
forecast = 210.725010 53.000000 228.250011 210.725010 0.000000 1.000000 11
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 12
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 1
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 2
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 3
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 4
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 5
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 6
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 7
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 8
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 9
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 10
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 11
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 12
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 1
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 2
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 3
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 4
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 5
forecast,trend,season,seq num = 210.725010 0.000000 1.000000 6

The forecast = Top-Items then showed all period buckets with qty = 210.725 for all periods present and future

EXPECTED BEHAVIOR
-----------------------
Users expected different values for each bucket

STEPS
-----------------------
The issue can be reproduced at will with the following steps:
1. Create forecast rule and demand history data
2. Create forecast set / forecast name
3. Run generate forecast - INCFIF concurrent request
4. review output and logs and see only same value for each forecast period

Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms