When the general stores configuration, supply and distribution networks have been defined, the majority of the logistical reorder problems can be solved automatically by means of computer programs, which use generally available statistical information. This information consists of dynamical data, such as issues and purchasing transactions, basic item and classification data, stores constants, and tuning factors for which an optimal value can be defined and which modify the working of the programs.

The purpose of this manual is to provide information about the system definition and setup. Only after complete setup, operations support can be given and programs, analysis and reports can be run.

The system can be setup as expert system for logistic operations support of the current real life situation on the production database, or to study the impact of business decisions on the current or expected future situations, by optimisation modelling and simulations on a test database.

This manual describes the functions related to basic data and tuning. Basic data consists of those data that may be regarded as elementary data whereas dynamic data is created by business events and depends on the elementary data. The system contains a complete set of basis functions and database tables (its own house holding).

Normally transactions will originate from business events of the production, or they can be generated by simulation. The function INU08 allows transactions of business events to be entered for simulation and modelling.

The following functions define the setup, the basic data and tuning.

The operations support functions (see OPERATIONS manual)

NOTE: The function ordering presents also the order in which data should be defined, to get the system operational. e.g. First general constants and rules in INU01, then control parameter group definitions INU02, then items in INU03, which may be linked to a parameter group defined in INU02.
 

The basic data can be entered and modified in screens and consist of mandatory data and non-mandatory details. For INU01 recommended defaults values are provided, which normally do not need to be adjusted. After tuning most of the set-up functions do not need any manual attention anymore. In stand-alone version or for simulation and modelling study purposes the functions INU03 - INU10 can be used to enter or change the basic data manually.

In production situation where the system is used as expert system to support another operational ERP (Enterprise Resource Planning) system like SAP, BAAN, ORACLE Logistics, the information in INU03 - INU10 is normally automatically maintained by interface to the operational ERP system. Only INU12 should be used during production, to support manual exceptions handling.

SETUP INU01

Only at setup. Recommended defaults are provided.

INTERFACE INU03 - INU10

Data automatically maintained by interface to production.

MANUAL INU02, INU11, INU12

Manual, to enter constraints and to handle exceptions.
 

Hierarchy

PARAMETER GROUPS

A number of management input, steering or control parameters are needed in the logistical calculations, such as target service-level, holding cost, fixed and variable order costs, absence cost, item review and entry time. Although DEFAULT values can be defined in the setup function INU01, which will serve most of the items, groups of items may still exist which should be treaded differently. e.g. They may have a higher order line cost, or a longer review time caused by a required inspection, or management may want to set a higher target service level.

For these groups of items it is not necessary to define the control parameters for each individual item. It is sufficient to enter the NOT-DEFAULT control parameters on a single parameter group record in INU02 and link this control parameter group record to the group of items in INU03.

In this way maintenance work is decreased and control is increased in a simple and easy way.

Only if an item needs its own UNIQUE control parameters, these control parameters should be defined for this item in INU03 /2.

The values of the control parameters can thus be specified on a number of levels. These levels define a hierarchy, which is followed when the value of a parameter is retrieved for use in a calculation.

HIERARCHY Item - Parameter Group - Default

If a constant is defined on item level, this value is taken, if it is not defined on item level, the value defined to the control parameter group is used, if this value is not defined, the overall default value defined in screen INU01 is used.

NOTE: Never enter a 0 if you want to leave a parameter undefined, the 0 will be taken as a value, only an empty, blank field denotes an undefined parameter.

If one only wants to use global values, define the default parameters in screen INU01 only: overall service level, holding cost, fixed and variable order costs, absence cost, review and entry time. All calculations will use these values. Definition hierarchies for the different parameters:
 

NOTE: As defaults for the item control parameter are NOT hard-copied into item records, but dynamically determined at calculation time, just a view high level parameters, whose values may be changed at any time, may control all the items.

The user sets the objectives and targets in INU01. Occasionally occuring non standard situations can be handled by defining additional parameter groups in INU02.

SIMULATION

Simulate by "real history replay method" (using real historical data with different calculated parameter sets) the impact of business decisions, analyse simulation results, trace exceptions, use optimisation modelling techniques to find optimal parameters and business policies

Definitions

A number of target or control parameters highly influence the results of the logistic calculations, some of these parameters must be defined (service level, holding cost, order cost, review time) while others are derived by statistics (demand statistics, lead times).

E.g. the economical order quantity EOQ depends on the total order cost and holding cost, the item cost price, stock demand or item consumption, and out of stock costs. The safety and buffer stock depends on the lead time and customer order statistics and requested service levels.

Service Level
Service level in percent %
100x (Number of successfully served requests/Total number of requests)

Holding Cost
Cost of holding the item in stock, calculated as percentage of the average value in cost price of the items hold in stock, by year.
An often by management imposed constraint is that the total stock value should not exceed a certain defined amount: the investment limit.

Fixed Order Cost
Fixed part of order or administrative cost. Cost of processing the order through accounting and invoicing, including order system costs.

Variable Order Cost
Variable part of order or order-line cost, involves handling, transport and insurance, it may include manufacturing costs of the product for various order sizes, distribution, inspection and entry.

Absence cost
Percentage, which is used to calculate the loss of profit to sales, when the item is not available to sales.
If the gross profit is defined by Qty * (Sales price- Cost price)
then Loss = BO Qty x Profit x (Absence cost/100)
An estimate of the yearly loss can be given as (1-Service Level/100) x Year turnover Qty x Profit x (Absence cost/100)
To sales it is important how much you are selling or not selling because the item is not available and the loss resulting from it. Absence cost is useful to calculate the cost of lost sales.
e.g. If a customer has to wait, for his order to be filled, the loss or absence cost is 0 %, if the customer does not wait and the customer order is lost to a competitor the absence cost is 100 %, if the customer is a regular client, but does not order anymore if an order can not be filled the absence cost or loss to sales may be higher then 100 %

Stock-out Cost
Expected real cost caused by stock-out per order cycle. This cost is of another type then the absence cost (see Operations Manual), and is more related to costs which occur when an item, cannot be issued for planned production or repair. The costs of lost production and costs of changing the work plan, and rescheduling the operations can be considerable. This cost is calculated per unit of the item.

Exception threshold
Exceptional large non-recurrent, or infrequent occurring large demanded quantities will lead to stock-outs and unsatisfied normal customer orders (see picture). If these exceptional large demands must equally well be satisfied from stock, this will lead to non-economical extremely high safety stock situations. To avoid these unwanted situations the demanded quantity is checked against the exception threshold level or break quantity. If the demanded quantity is above the exception threshold level, it is turned into a special order or so-called Direct Out. Demanded quantities below the exception threshold level are considered to be Normal Demand, and are normally served from stock.

Pattern Factor A system seasonal pattern flag is set, if the average deviation of the consumption over seasonal corresponding periods divided by the average deviation over all the periods is smaller then a threshold defined in the setup function INU01.
Avg (Dev (P Seasonal)) / Dev (P) compensated for trend, calculated over a minimum of 2 years.

Trend Years A system trend pattern flag is set, if over the last years there existed a steady increase or decrease: the trend gradient value was large, with a small variation of this gradient.

Lumpiness Factor A lumpy demand system flag is set if the normal stock demand Q calculated by month is very variable, as defined by the ratio of the deviation to the average demand sampled in month periods.
Dev (P) / Avg (P) > in setup INU01 defined factor.
The lumpy or erratic demand may be perpetual, composed of periods of little or no demand followed by periods of high demand.

  SEASONAL PATTERN Variation over seasonal related periods is much smaller then the overall average variation.
  TREND Average value of the differences between the years is large, and the deviation over the differences is small.
  LUMPY Variation over the periods of consumption is much larger then the average value.

Main menu

In order to access the setup and basic data functions one has to select the role LOGIC main menu

INU01 /1 -SYSTEM SETUP CONSTANTS, DEFAULTS AND RULES

FUNCTION AIM

The purpose of this function is to define general system settings, it allows to define DEFAULTS for the target or control parameters. Such as target service level, holding cost, fixed and variable order cost, stock-out cost, time windows, management objectives, rules and constraints.

FIELDS

Target One can define an overall target service level or one can choose for the minimisation of the total cost, in the case of minimum total cost, the service level for each item is a result of the minimum total cost calculation.

Investment Limit This constraint sets a limit to the total average amount in item cost price, of inventory that may be carried by the stores.

Time Window A summary of historical events like number of orders, total quantity and amount, for a time window or period of history is kept on the inventory item, supplier and customer records and used as input to the ABC analysis. Different time windows and ABC threshold percentages can be defined.

HOW TO USE THE FUNCTION

Set the parameters to the desired values. By pressing <TAB> buttons on the left one can reach other tab pages. Not all fields are mandatory.

INU01 /2 - COST BALANCE FACTORS

DESCRIPTION

The purpose of this screen is to define the general cost balance factors. In general there is a cost to order/transfer an item, which is composed of 2 components: the order/transfer setup cost and the order/transfer line cost for each additional item. It costs to have or possess the item, the so-called holding cost. It may also cost if an item is not directly available from stock when there is a material request for this item, the so-called stock out cost. Finally there is a certain cost related factor if the item is hold in stock but there is no demand for this item.

FIELDS

Order/Transfer Setup cost  Fixed part or administrative part of the order or transfer cost.
Order/Transfer Line cost  Additional cost for each order line or transfer line.

The total Order/Transfer cost can be defined as order setup cost + line cost x number of lines.

Holding cost  Cost of possession as percentage of the average item value per year.
Stock out cost Extra cost if the item is not issued on demand.
Salvage value If there is no demand for an item, it still may have a certain salvage value, which represents the reclaimable value, which may be recovered if the item is returned to the supplier or sold out under discount.

INU01 /3 - TIME FACTORS

DESCRIPTION

The purpose of this screen is to define the general time factors. Some time factors can be obtained from the system history, like time between ordering and reception, while other time factors have to be estimated and entered by the user, like administrative time of reviewing the re-order propositions, inspection and entry of received ordered items into stock.

FIELDS

Order time
Time elapsed between the order sending and reception of the ordered material.

Extra time
Estimate sum of all the administrative components that have to be added to the order time, like re-order propositions review time, reception and inspection time, time to enter the item in stock.

Lead time
The total time elapsed in the reorder process between the moment a replenishment need is detected and an order proposal is generated, and the availability of the item in stock to sales of the re-ordered item.

The purpose of this screen is to define general order and transfer constraints. These constraints are logical and generally defined by the management to avoid unwanted situations. e.g. It may be economical to order 3 years of material, but management wants to stock not more then 0.5 year of expected consumption: a lower order frequency limit constraint of 2 will restrict the EOQ quantity of 3 years (frequency 0.33 lines per year) to 0.5 year of expected consumption.

ORDER/TRANSFER CONSTRAINTS

Upper limit Limits the number of orders per year, or inversely poses a lower limit to order quantities as compared with the expected year consumption.

Lower limit Lower order frequency limit. This condition poses an upper limit to the quantity ordered per year.

Lead time limit A long lead time and small economical order quantity EOQ may lead to multiple orders in the lead time. This constraint limits the number of orders that may simultaneously exist. It relates the order quantity to the consumption in the lead time, imposing long lead times to result in a larger order quantities.

CONSOLIDATION

Ratio ROP Ratio of the pull down quantity and the reorder point
Time Days of daily consumption before reordering

CONSOLIDATION

The consolidation process aims to synchronise ordering of items for the same supplier and contract, and results in orders with many order lines. It works by pulling other items into reordering condition, when one item is normally to be reordered for a given supplier and contract. In order to get other items into reordering the reorder points are artificially increased.

A number of factors can be defined.

If more factors are defined they will work together: if one condition is not met, but another is satisfied, the item will be pulled down into reordering. The consolidation process is stopped after the following conditions are reached

Minimum value for the order is reached
Maximum number of order lines for the order is reached.
 
 

A number of limits to the increase of reorder point can be defined, each limit having its own benefits.

Ratio ROP (OH-ROP) / ROP < limiting condition

The increase limit by ratio of the pull down quantity = (quantity in stock OH - quantity reorder point ROP) and the reorder point. This condition has the disadvantage that there is no control on the extra costs.

Time (OH-ROP)/Daily Consumption < limiting condition

The increase limit by the pull down time = (quantity in stock OH - quantity reorder point ROP) / Daily Average Consumption. This has the advantage that items that will soon normally come into re-ordering, are re-ordered grouped together.

INU01 /5 - SAFETY STOCK MODULATION MATRIX

DESCRIPTION

The safety stock quantity is calculated as a function of the service level and item demand history.
Factors in this calculation are variability in expected demand quantity and number of demands and lead time, but there are NO economical cost price conditions imposed.
This matrix allows modulation of the safety stock quantities by multipliers in an economical way, by obtaining higher service levels at relatively low extra holding costs. If the factors are taken larger then 1, the safety stock quantities will be larger, and higher service levels, then set by the target service level definition will be obtained.
The matrix horizontal axis represents the ABC classification by frequency or number of demands
The matrix vertical axis represents the ABC classification by safety stock value

MATRIX FIELDS

Coefficient AA - (item demand ABC = A, safety stock value ABC = A)

Coefficient AB - (item demand ABC = A, safety stock value ABC = B)

INU01 /6 - Order Time Statistics

DESCRIPTION

The purpose of this screen is to define or change order time statistics calculation factors.
The parameters defined in this function will be used to calculate order delays for suppliers and supplied items.

FIELDS
 

Time Window	Time window in years of transaction history used.
Filter Minimum	Order delays below this number will not be used in the calculation.
Filter Maximum	Order delays above this number will not be used in the calculation.
Filter Exception Factor	Data larger then the average order time times this factor will not be used in the calculation.
Weighting scheme	Weighting scheme used, to calculate the average order time and deviation.
Weighting factor	Smoothing factor used in the weighting scheme to give more weight to recent transactions
Confidence number	Below this number, there are insufficient transactions for the item, and a combination of supplier and item statistics is used.

DATA ERRORS and ACCURACY

New order delays are calculated based on order transactions. From these transactions the average and deviation are calculated. However, the results of this calculation are highly sensitive to data errors, abnormalities, and historical time interval used. Different filters can be defined to avoid abnormalities and enhance the accuracy of the statistics.

FILTERS

Time window historical data
This filter limits the historical transactions used to a pre-defined number of years
e.g. historical data > last 5 years

Fools filter
The fools filter or data entry errors filter, filters the logical abnormalities out
e.g. order delays > 365 days and order delays < 0 due to data entry errors are filtered out

Statistical exceptions filter
The statistical exceptions filter, filters the statistical abnormal events out for order delays > Factor x statistical average.
e.g. order delays > 4 x statistical average

STATISTICS

After filtering the data, the order time statistics is calculated

Therefore historical time dependent weighting, based on the number of years should increase the accuracy

Unit: W=1, No weighting

Linear: W = greatest ((Factor-Y), 0)

Reciprocal: W = 1/(1+greatest ((Factor xY), 0))

Reciprocal: W = exp (- Factor xY)

CONFIDENCE

In case of item introduction or change of supplier there are often insufficient data to determine accurate order times statistics by item with confidence. Therefore the algorithm calculates in 2 steps.

Note: at T=1 D (Item) by definition = 0 and D (Item) is taken as D (Supplier)

INU01 /7 - Consumption Statistics

DESCRIPTION

The purpose of this screen is to define single demands statistics factors like average expected quantity per demand and exceptional threshold level or break quantity formula factors.

FIELDS
 

Time Window	Time window in years of transaction history used.
Weighting scheme and factor	Single demanded statistics may be calculated using time weighting. Weighting focuses the calculation on more recent issues.
Exception Factor	The exception factor N determines the statistical exception threshold level or break quantity E = A + N x greatest (A, D).  The exception level may be constraint by economical factors.
Maximal Value Constraint	If the value of the exception > maximal value, the factor N is shifted to a smaller value, this smaller value however has a lower limit, defined by the minimum factor.
Minimal Factor	Minimum value N may take under Maximal Value Constraint.
Minimal Value Constraint	If the value of the exception < minimum value, the factor N is shifted to a larger value, this larger value is restricted by the maximum factor, imposing the upper limit.
Maximal Factor	Maximum value N may take under Minimal Value Constraint.

EXCEPTIONS

The exceptional threshold level or break quantity is defined to avoid stock-outs, which will be caused by serving non-recurrent or infrequent occurring exceptional large demands from stock. If infrequent occurring exceptional large demands are included into the statistical demand used to calculate stock replenishment parameters, this also will lead to high safety stock values.
Careful definition of an exceptional threshold level quantity, above which a demanded quantity will be served by a special order or direct-out, instead of being served from stock, will lead to high service levels in combination with low stock values.
A definition of statistical infrequent large demands can be formulated by the demand statistics parameters average and deviation. Requests for a quantity E larger then the average A plus N standard deviations D will occur only with very low probability.
 

Sometimes D is very small or even zero e.g. when for a large number of demands always the same quantity is requested, thus a more practical definition is given by

E > A + N x greatest (A, D).

The above definition, does not include any price considerations. Practically one wants expensive items to be served by direct-out quicker (less rare in statistical sense) then very cheap items.
Economical conditions can be imposed by shifting N to a smaller value (less rare occurrence of exceptional demand) if the cost of the break quantity E is very large.
This constraint will lead to quicker ordering by direct-out of expensive demands.
To avoid a very small value of N, a minimum can be defined in INU01
 
 

The same reasoning like above, but now for cheap items, leads to shifting the N to a larger value if the cost corresponding to E is very small. This constraint leads to nearly always serving from stock of cheap item demands even if they very seldom occur. To avoid a very large value of N a maximum can be defined.

If there is sufficient quantityin stock, an exceptional large demand may be served in the normal way from stock.
This is the case if after serving at least one reorder point + average daily consumption is left.
 

In order to avoid manual checking this condition for all daily exceptions, a so-called stock dynamical exception threshold level can be calculated such that after serving the exceptional demand at least

(Stock Dynamic Factor) x (the reorder point + daily consumption)

should be left in stock for normal serving and reordering

NOTE 1: The stock dynamic break quantity, is dependent on the situation of the stock, therefore it should be daily recalculated.

NOTE 2: The statistical quantity used for forecasting of normal demand depends on the statistical exception level, or user defined as constraint direct-out level but NOT on the stock dynamical level.

INU01 /8 - Tuning Factors

DESCRIPTION

The purpose of this screen is to set or modify a number of factors, which influence the logistics in a global way. For instance all economic order quantities are multiplied with a tuning factor, whose value is normally set to 1, assigning any other value will increase or decrease all order quantities. In this way the user can control and globally fine-tune the logistics to his own judgement, for instance set all order quantities to a global larger or smaller value.

FIELDS
 
 

Order Qty Tuning Factor	Globally modifies the economical order quantities. The order stock quantity is multiplied with the Order Qty Tuning Factor NOTE: This factor is applied BEFORE the order constraints (see INU01/4) are applied.
Marginal Correction Factor	This term is important for items which are perishable (e.g. ink) or have a short and defined shelf life, or the demand for them is a one-time event (e.g. agendas). As the demand is variable and covered by a single order (e.g. when the expected consumption in the shelf life < EOQ) there may be a loss calculated by the salvage value for unsold product, as well as a lost profit if the demand exceeds the ordered quantity. The order quantity that will maximise profit can be calculated from the expected demand plus a correction depending on the deviation of the demand, the cost price, the sales price and the salvage value. This correction is multiplied with the Marginal Correction Factor.
SL Lower limit	Target service level is defined over the total time the item is served from stock. This time consists of two parts: The time in which the stock above the reorder point quantity will be consumed, and the time during which reordering takes place. If the first component is relatively large, this may by definition lead to a low service level to be satisfied during the lead time. To ensure that also in the lead time there is sufficient service, a minimum service level can be defined here.

LOWER LIMIT

The service level for an item is defined as the ratio of the number of requests that can be immediately satisfied from stock and the total number of requests.

Service Level in % = 100 x Number of served requests/ Total number of requests for material

The target service level is defined over the total time the item is kept in stock. Normally there is only a possibility on stock-out and backorders after the quantity in stock has reached the reorder point. For the time the stock is above the reorder point, almost 100 % service level is ensured.

If the order serve time is very large compared with the reorder time, this will by definition lead to a low service level in the lead time, to reach the overall defined target service level.

ROP SS Tuning Factor
This term is important if one globally wants to modify the calculated safety stock quantities. The safety stock quantity is multiplied with the ROP SS Tuning Factor.
This factor is directly applied to the safety stock component

FOC SS tuning factor Allows to fine tune the FOC safety stock in maximum or fill-up-to level. This factor is directly applied to the safety stock component

For a number of years and a minimum number of demands in these years the trend, pattern and lumpiness are derived and corresponding flags are set on the stock item (see INU04 /2)
See DEFINITIONS - the sensitivity of calculation can be increased or decreased by the
Trend Factor, Pattern Factor, Lumpiness Factor, set stock item Flag or Sign if
                       Sign (dQ) equal, dQ/Q > Trend Factor
                       Dev (Qs) / Dev (Q) < Pattern Factor
                       Dev (Q) / Avg (Q) > Lumpiness Factor

Average Inventory Factor Normally the quantity in stock fluctuates very much as result of the replenishment cycles, and the average quantity kept in stock is important management information, it is very convenient to keep a time weighted exponential smoothed average of the qty in stock S(Q) on the stock record for reporting purposes:
t = system time - last calculation date
f = t / ( t + 365 x ASF)
S (Q) = f x Q + (1-f) x S(Q)

NOTE: The calculation is somewhat insensitive to the period between re-calculations. The algorithm can be re-run on the same day (no action), or re-run in intervals of days, weeks or every month and still results in about same smoothing which depends on the factor ASF but not on the period between different runs.

Maximum History Limit For data maintenance, when defined, older history will be removed

System Calculation Date If this field is defined its value is used in the calculations instead of the SYSTEM DATE. It should normally NOT be defined.

INU02 - Parameter Groups

FUNCTION AIM

The purpose of this screen is to define parameters groups. The parameters target service level, cost of holding, order line cost, cost of absence can be specified. If the parameters are not defined on the item level (INU03) and a parameter group is associated with the item, the parameters of the parameter group will be taken instead of the defaults (INU01).

FIELDS

Code, Description      Unique identification code, description of the parameter group
Service-level, Extra time, Cost of order, Cost of holding, Cost of absence, Salvage value
See HIERARCHY and DEFINITIONS

HOW TO USE THE FUNCTION

Enter the code and group description. The parameters do not need all to be defined, if the item specific or default values are to be used for the item. When pressing <DELETE RECORD> a warning will be used when the group code is in use, the group code will be removed from the items using the class when <COMMIT> is pressed.

INU03 /1 - Item Basic Data

FUNCTION AIM

The purpose of this screen is to define or modify item definition and classification details.

FIELDS
 

Item code, Description	Unique identification code, item description.
Control	Item responsibility center/party or person, Manager
Unit	Unit of Measure UM, Basic Unit - this is the accounting unit.
Status	Identifies events in the life cycle. E.g. new/ active / not active.
Cost	Average unit cost: this cost is used to evaluate the stock value and calculate the economic order quantity.
Price	Sales price, which is charged when the item is sold.
Stock Unit	Default stocking unit
Conversion	Conversion factor between UM and stocking unit
Principal Store	Main store, which is replenished for the item in a single re-order point many stores configuration. Purchasing is only re-ordering the item for this store, this store however may feed many satellite stores by inter warehouse transfer
Shelf-life	The number of days the item can be held in stock before it will not be fit for consumption anymore, e.g. it will deteriorate or corrode after this date.
Estimate	Initial year consumption quantity estimate for a new item. On introduction of a new item, there is no statistical data, the year estimate will be used to calculate the initial parameters.
Introduction	Introduction /creation /item definition data.
Order time	Time needed for the supplier to deliver the item.
New Version	Date on which a new version of the item is due.
Class	Identifies the physical class
Index 1	Free user definable index 1 e.g. Type identifies the logical class or use as spare part or consumable or cleaning use.
Index 2	Free user definable index 2
Principal Supplier	Supplier at which the item normally will be reordered. Lead times statistics are calculated and the contract is valid for this supplier.
Contract	Current valid contract.
Order Minimum	Minimum order quantity. This is an order condition imposed by the principal supplier.
Order Multiple	Incremental or order multiple quantity. This is an order condition imposed by the principal supplier.
Remark	Free user definable remark or text. The <LIST VALUES> button will display a list of remarks for the current item control responsible or manager, for retrieving (enter query or interrogation mode) or definition.
<IMAGE>	When pressed the item image is displayed (see INU04 /1)
Description	Item text

HOW TO USE THE FUNCTION

Use <CREATE RECORD> to enter an item. Item properties can be selected using help screens by pressing <LISTVAL> on the item class field.
By pressing <TAB> on the left of the screen, the next TAB page will display a number of target or control parameters which are used by the logistical calculations.

INU03 /2 - Item Control Parameters

DESCRIPTION

This page allows to link a parameter group to the item, and/or to define specific values for the target or steering parameters to the item. When a parameter group is defined for the item its target parameters are shown. The DEFAULT values defined in INU01 are shown as well.

FIELDS
 

Parameter Group	A parameter group can be linked to the item, providing specific control parameters specific to the item.
Service Level, Order Cost, Holding Cost, Absence Cost, Salvage Value	see Definitions
Extra time	Review/Entry time: Total time needed to review and confirm, and to enter in stock of the ordered item. This time may include the extra time needed for assemblage or inspection, before the item becomes available.
Stock-out Cost	Estimated cost of stock-out, usually related to costs which occur when a planned item, cannot be issued, or cost of production stop if the item is used as spare part. Costs of changing the work-plan, and re-scheduling the operations, or production failure can be considerable. NOTE: This cost is defined per unit of the item.

INU04 /1 - Stock Administration

FUNCTION AIM

The purpose of this screen is to display, define or modify stock definition details.

FIELDS
 

Store, Item	Store, Item.
Control	Item responsibility center/party or person, Manager
Unit	Unit of Measure UM, Basic Unit - this is the accounting unit.
Price	Sales price charged when the item is sold.
Status	Identifies events in the life cycle.
Stock Qty	Total quantity in stock, possible in different stocking locations or Binloc
Stock-out Date	For current stock-out, the date on which qty became 0
On Order	Qty on order
On Reserved	Qty of currently not satisfied requests
Stock Unit	Default stocking unit
Conversion	Conversion factor between UM and stock unit
Minimum Level	Calculated or by constraint imposed logical minimum.
Maximum Stock	Maximum quantity that can be stocked or that one logical wants to hold in stock.
Lot Qty	Calculated or by constraint imposed logical set size or quantity.
Direct Out	User defined break quantity at which a material request normally will not be served from stock but turned into a special order.
Supplier	Default supplier from which the item normally will be reordered.
Main Store	Supplying store if current store is a satellite or transfer store
Order Time	Statistical order time of item by supplier

DERIVED BY DEFINITIONS ACCORDING HIERARCHY:

Item (INU03)- Parameter Group (INU02) - Default value (INU01/1,2)
Service Level Service level in %

Holding Cost, Stock-out Cost, Salvage Value

Costs, calculated in the system currency

DERIVED BY DEFINITIONS ACCORDING HIERARCHY:

Item (INU03) - Parameter Group (INU02) - Default value (INU01/1,2)
Item, Group Transfer Channel (INU05/2) - Default value (INU01/3)
Extra Time Extra time component for orders and transfers
Setup Cost, Line Cost In system currency for orders and transfers
NOTE: By pressing the <IMAGE> button, an image of the item is displayed.

FIELDS
 

CO	Number of customer orders
Qty CO	Total quantity CO turnover (issued - returned)
Turnover Value	Total amount CO
Benefit	Gross benefit = (Sales Price-Cost Price) x Qty CO
Direct Outs	Issues not from stock but as special orders directly from supplier
Backorders	Not immediately served demands
Returns	Returned
Transfers Out	Inter warehouse transfers out
Transfers In	Inter warehouse transfers in
Trend, Pattern, Lumpy	Sign and flags, see DEFINITIONS
First/last Transaction Date	Date first/last transaction
Total Number CO	All CO ever
Total Qty CO	All Qty CO issued ever.
Qty/Issue	Average quantity per demand
Number CO/Day	Average number of CO per day
Qty/Day	Average quantity demanded per day
Statistical Exception Qty	Statistical Exception Threshold
Break Qty	Break Qty or Exceptional Qty Threshold level

INU04 /3 - Demand Forecast

FORECAST NORMAL DEMAND
Expected total demanded quantity from stock, including issues, transfers and returns. This quantity is important for stock re-order parameters.

FORECAST EXCEPTIONAL DEMAND
Expected total ordered quantity from direct orders, delivered directly to the users, this quantity is important for contracts and non-stocked items.
 
 

Method	Forecasting method, which produced the best results - smallest average error or squared error - over the last (iterations) historical periods.
Forecast	Short forecast over by setup (MAU01) defined
+/-	Standard deviation or Forecast Error
Year Forecast	Forecast over year period - used for contracts and economical order quantity
Iteration	Number of historical periods over which the forecast method was tested (usually 12) defined in the setup (MAU01), sometimes less due to lack of history.
Bias	Average signed error, denoting systematical under or overestimation of forecast
Planning	User defined or by planning defined. IF DEFINED overrules the statistical year forecast. Value will be erased if <Date expires.

< End of plan date, if DEFINED this date will when EXPIRED, ERASE the planning and its own value setting planning and date to undefined. Note: if UNDEFINED the planning definition will stay until by user action modified or erased.

INU04 /4 - Parameters

 

Policy	Re-order policy: ROP - Reorder at reorder point FOC - Reorder at fixed order date MIN - Reorder at user-defined minimum to user-defined maximum
Stock Reserve	Average expected consumption in lead time
Safety Stock	Added extra term to overcome variations (larger than expected) consumption in the lead time
Reorder Point	Point (quantity) which triggers re-ordering: if available quantity in stock drops below this quantity, a reorder proposal is generated.
EOQ	Economic Order Quantity see DEFINITIONS
Consolidated ROP	ROP be temporary increased by consolidation conditions (see INU01 order constraints)
Order Qty	By logical constraints restricted EOQ
FOC interval	Review period between FOC dates
FOC Date	Next date at which an FOC order is triggered. Automatically maintained by the system.
FOC Trigger Level	FOC order line created on FOC date when quantity in stock below this level
FOC Max Level	Order line quantity calculated to fill up to this level
Transfer ROP	Transfer from main store, if quantity drops below this level
Transfer OQ	Transfer order quantity
Con Order Time	Time to calculated a target arrival date on the order receipt, the same for all lines on the order (same supplier, contract)
Dynamic Break	Break qty also taking actual stock situation in account

INU05 /1 - Stores

FUNCTION AIM

The purpose of this function is to define the different stores, store types, network definitions, transfer channels, stock locations, quantities and units.

FIELDS

Store Store code and name
Type Stores type:
0 = Virtual store is a logical or administrative store
1 = Main store is replenished by suppliers
2 = Transit or Outlet stores are re-filled by transfers from others stores

INU05 /2 - Network

DESCRIPTION

This screen allows to define the transfer costs and times for transfers through the various transfer channels. The transfer cost and time can be specifically defined for a given item or parameter group, or can be more generally defined for a given transfer channel between a receiving store and a master - or providing store.

FIELDS
 

Store	Satellite or detail store receiving the material.
Master	Master or main store supplying the material.
Priority	Order in which a transfer of items is attempted
Parameter Group	Group definition (see INU02), transfer channel is linked to the items of this parameter group
Item	Specific item for which transfer channel is defined
Time	Transfer time
Transfer Line Cost	Transfer line cost

INU05 /3 - Binlocs

BINLOCS or STORES LOCATIONS
 
 

Store	Store
Binloc	Store location code
Item	Item
Count Date	Date last cycle count (control of quantity)
Type	User-defined location type, like cooled, or heavy objects
X, Y, Z	Positional parameters, like path or alley, level or shelf, case
B	Blocked for picking, denotes that the location is temporarily blocked. e.g. for counting purpose
P	Priority or pick first rule, e.g. caused by shelf-life
S	Sequence number, calculated to optimise picking route

QUANTITIES
 

Item, Qty, Unit	Item quantity in unit of measure
Lot Id, Lot Unit	Lot oritem identification (tracked items), stock unit
Insert	Put away date
Update	Last picking or change date
Count	Last cycle count date

INU05 /4 - Units

DESCRIPTION

This screen allows to define the units. Units can be used a basic unit or unit of measure, like kg, meter, litre, piece or more general like box, pallet, bottle

FIELDS

Unit Unique unit code and description
UM Checked if unit of measure
Conversion Relation to other units

INU06 /1 - Suppliers

FUNCTION AIM

The purpose of this function is to define and modify suppliers.

FIELDS
 

Supplier	Unique supplier code and short description
Index 1, Index 2	User-definable indexes e.g. to denote type or classification of supplier
Status	Status
Address, City, Country	Address, City, Country
Contact	Contact person(s)
Phone, Fax	Phone number, Fax number
E-Mail	E-Mail - pressing <E-mail> starts Email
WWW-site	Web-site - pressing <WWW-site> starts the internet-browser
Supply time	Supply time
Comment	Short comment / remark
Items	Number of items currently supplied
Orders	Number received orders historical time window
Amount	Total amount received orders
Order time	Average order time
ABC	ABC Coefficients

when <E-mail> pressed the SYSTEM INTEGRATED Email is started ...

DESCRIPTION

This function allows to display, define and modify supplies, and historical supply conditions. Normally it is automatically maintained by interface, a historical record is created each time the active supply conditions change.

FIELDS

DESCRIPTION

This screen allows to define the currencies and rates.

FIELDS

Currency Currency type and description
Rate Exchange rate
Update Last modification date

INU07 - Customers

FUNCTION AIM

The purpose of this screen is to display, define, alter or delete customers.

FIELDS
 

Code	Client code
First Name	First name
Last Name	Last name
Phone	Phone
E@	E-mail
Company	Company, division or organisation
Office	Office location
ABC statistics	Number and total amount issued in history time window, and ABC classifications

NOTE: if the <Code>, <Division>, <Amount> and <Number> header buttons are pressed, the order in which the data in the screen appears will change accordingly

INU08 - Transactions

FUNCTION AIM

This function allows to display, define, alter or delete transactions.

HOW TO USE THE FUNCTION
Transactions should ONLY for simulation purposes in test-environment be entered or changed. The transaction screen is wider then the window, by using the <SCROLL-BAR> the screen will scroll and display the next fields. In UPDATE mode, it allows to enter and modify transactions for existing items. Entered positive quantity transactions will be treated as issues with default type I. Negative quantities will be treated as returns with default type R. In QUERY mode the screen can only be queried and will display full information about the selected items and/or dates and business events.

FIELDS

EXAMPLES

Management information
Root cause analyses (maintenance, work repair cause analysis)
    Observation - what happened
    Direct Cause - cause
    Indirect or root cause - 2e cause
Analytical accounting
    Cost type/reason - type of - what happened
    Cost budget - who pays
    Cost originator - who cause
Reasons for Adjustments or Corrections of stock
    Lost/missing when picking
    Lost/missing when counting
    Found when picking
    Found when counting
    Quality due to store (e.g. corrosion, dried out when keeping too much too long)
    Quality due to supplier (e.g. faulty technique)
    Damage
Reasons for Returns
    Wrong quantity due to requestor
    Wrong quantity due to store
    Wrong material due to requestor
    Wrong material due to store
    Quality
    Damage
    Delivered too late

TRANSACTION TYPES

The pop-up help screen or list of values shown below is available wherever an existing transaction type has to be entered e.g. for reporting or display of information or extraction of data to EXCEL.

It is an example of the many help-screens and lists of values available in the system

The following transaction types are supported

I = Issue of material by client order
D = Direct Out of client order by special purchase order
R = Return of material by client order
A = Adjustment or correction stock quantity
P = Sales Price Adjustment
C = Cost Price Adjustment
E = Entry of ordered material
T = Out going inter warehouse transfer
U = Incoming inter warehouse transfer
N = Not accepted order reception, return to Supplier
S = Stock-Out time
W = Work Order

INU09 - Control Centers, Managers

FUNCTION AIM

The purpose of this screen is to display, define, alter or delete item control or responsibility centers.

FIELD

Code Responsible party, person or manager code
Description Name or description
E-mail E-mail address
Item Number of items

MAU01 /1 - Forecasting Setup

FUNCTION AIM

The purpose of this screen is to define the initial forecasting set-up. These parameters should normally not be changed anymore.

FIELDS
 

Period Grid size	The length of the time interval for which the historical demand is sampled and the forecast is calculated.
Long Forecast	Number ofperiods long forecast to calculate re-order quantities and contract conditions
Short Forecast	Number ofperiods short forecast to calculate re-order triggering conditions
Maximum Number Iterations	The number of periods in the past for which the real historical consumption and the forecasts are compared. The forecasting methods are evaluated by testing the fit to the real historical quantities
Minimum Number Iterations	Minimum number of tests for a forecast formula, if there is not enough history for a formula to perform the specified minimum, the formula is bypassed.
Weighting Scheme	A weighting scheme may be applied to the tests, aimed at increase of the accuracy, giving more significance to the most recent tests or deviations between forecasts and historical periods
Weighting Factor	Factor applied in the weighting scheme
Best Fit Selection Criterion	Optimal method selection as best fit to: 1 = Least average squared error <E2> 2 = Least average absolute error <|E|> 0 = No selection, fixed method

HOW TO USE THE FUNCTION

This function displays information about the forecasting set-up. By pressing <TAB> buttons one can reach the next pages.

DESCRIPTION

Whatever would have worked best in the recent past should be used to predict the future.
A large number of different forecasting strategies is used simultaneously to forecast a number of recent historical periods of demand. Each formula is measured according to the variance from what actually happened in the know past, and the formula which worked most accurately, is selected and used to predict the future.

BENEFITS

The forecasting uses a multi-method procedure simultaneously calculating a large number of forecasts for a number of historical periods, using a large set of different forecast methods or formula (defined in MAU01/2), and subsequently selecting the method best fitting to the historical consumption for each individual item. By dynamically choosing the method giving the best fit to the historical consumption one ensures that:

The statistical used consumption includes customer returns, and a weighted fraction of the exceptional demanded quantities. Set-up allows to complete/partial/not including a fraction of the exceptional demand by the Exception Filter.

The exceptional demand is defined as demand for quantity larger then the break quantity calculated by parameters defined in INU01 or user defined as constraint direct-out level INU04

Although the number of exceptional large demands is only a fraction of the total number of demands, the total exceptional quantity may outweigh the total normal demanded quantity. Sometimes it is possible to satisfy an exceptional large quantity from stock without endangering the serving of the normal demands. Sometimes a fraction of the exceptional large demanded quantity is served from stock and the balance is served by direct out. Including a fraction of the exceptional demand will normally reduce the forecast error. Exception Filter setting:

-1 Take the total exceptional quantity for statistics
0 Take as fraction the exceptional threshold level
1 Decrease the fraction taken for statistics if the quantity
 

CALCULATION OF A FORECAST

For every forecast method a number of coefficients Ci can be defined in RPU05, such that a forecast for the next period can be calculated, expressed in the coefficients and the periods.

Forecast F by method m for the next period P

Fm (P) = Sum (Cm(i) x Pi )

The setup provides a number of standard methods and coefficients.
 
 

The period trend in the example, is calculated as the last period, plus the difference from the previous period: P1 + (P1 - P2) = 2 x P1 + -1 x P2

NOTE: Due to negative coefficients and/or user returns the forecast may become negative.

if the switch Allow negative Forecasts = Y. Allowing negative values of the forecast will make the test for the best fitting method more critical (One 0 is much alike to another 0, if no negative values are allowed).

CHOOSING THE BEST METHOD

The different forecast methods can be tested, how well they would have produced a forecast of the last N historical periods P1 ... Pn or Maximum Number Iterations N, by calculating the historical consumption R(P1 )... R(Pn) and forecasts for each forecast method m Fm(P1 ) ... Fm(Pn) over the last N periods and calculating the difference or fit of each method to the historical consumption.

The number of iterations must be chosen sufficiently large to reduce forecast error. It does so because it reduces formula selection on random change. A formula may be accurate once by accident, but may not stand the test of a large number of iterations. A too large number of iterations may decrease the accuracy, because it focuses the test on the fit to history of the far past.

The recommended default for the setup parameter Maximum Number Iterations is 12.

If there is not enough history for a formula to perform the specified number of iterations, it may run with a lower number of iterations. If the number of iterations is less then 3 it is bypassed altogether.

To test each method, the differences E between the real R consumption and the forecast F for each method m are calculated: Em(P1) = Fm(P1)-R(P1) ... Em(Pn) = Fm(Pn)-R(Pn)

For each method the average error, average absolute error and average squared error are defined as

<Em> = 1/(N-1) x Sum ( Em(P1) ... Em(Pn) )

<|Em|> = 1/(N-1) x Sum ( |Em(P1)| ... |Em(Pn)| )

<Em2> = 1/(N-1) x Sum ( Em(P1) 2 ... Em(Pn) 2 )

Where N is the number of actual iterations.

For random errors <Em> is near to zero. A large value of <Em> is caused by a systematic offset or trend between the predicted and real values. The value of <Em> is taken as TREND factor Tm for the method m, one advantage of adding Tm to the forecasts Fm is that all errors are reduced and the fit is improved.

the predictions Fm(P1)+Tm ... Fm(Pn)+Tm are compared with R(P1) ... R(Pn), and <|Em|>, <Em2> are calculated ( <Em>= 0 is per definition )

The best to fitting method or the method with the least error can now be selected as the method with the smallest value of <Em2> or <|Em|>. For this method we can obtain for the next period an estimate Fm(P0) with error <|Em|> and trend Tm.

Set-up allows both the fit to <E2> or <|E|> as best method selection criterion.

(<E2> is general accepted as best criterion, sometimes <|E|> is preferred for noisy data...)

The set-up also allows a weighting scheme to be used, to focus the method fit on the most recent periods:

<E2> = N/(N-1) x Sum W(i)*E2(i) / Sum W(i) where E2(1), is the squared error of the most recent period, E2(2) the next, etc., (same for weighted <|E|>)

FORECAST SCHEDULE

If a schedule of one forecast period is used, for the recommended Periods/Year = 4 setting
an accurate estimate for the NEXT 3 MONTHS is obtained for:

YEAR ESTIMATE = 4 x FORECAST + TREND x W(I) + Deviation Factor x 2 x DEVIATION

with a lower limit of Deviation Factor x 2 x DEVIATION

The lower limit is imposed because the forecast can never be more accurate then the deviation or error itself. The total error of the year estimate as square root (4 independent errors) = 2 error of one forecasted period. The factor W(I) depends on the number of real iterations (between Minimum Number Iterations and Maximum Number Iterations).

For seasonal items a full year schedule of the next 4 periods is calculated and 4 x FORECAST in the formula is replaced by FORECAST (P1)+(P2)+(P3)+(P4)

Note: Full year forecast schedules can be obtained if the setting Forecast Schedule Periods is set to Periods/Year.
 
 

DESCRIPTION

Forecasting allows any number of methods, to be defined. The purpose of this screen is to define, de-activate, alter or delete Forecast Methods.

FIELDS

Id	Unique code of forecasting method.
Description	Short description of forecasting method.
S	Status: A = Active, C = Not Active.
T	Type: F = Forecasting method, L = Life cycle pattern
P	Flag: P = Seasonal pattern, N = No seasonal pattern
Index	Coefficient number.
Coefficient	Coefficient Factor.

HOW TO USE THE FUNCTION

New forecast methods can be entered by pressing <CREATE RECORD>.

In the next block the coefficients for the entered method can be defined or modified. A forecast method may be removed by pressing <DELETE RECORD> when the method is in use a warning is issued. When <COMMIT> is pressed after this warning, the method definition will also be erased from the items using the deleted method.

MAU01 /3 - Mathematical Tables

THE MATHEMATICAL TABLES

FIELDS
 

Cumulative	Cumulative Probability P(Z) in % Level as approximation of the area under a Standardised Normal Distribution. This table will be used in calculations if the Real Distribution is not available (see Operations Manual)
Factor	Z Factor -Factor relating the service level and Standardised Normal distribution and Unit Normal Loss Integral
Area	Cumulative Probability P(Z)
Unit Normal Loss	E(Z) Factor relating the Unit Normal Loss Integral
Function	F(Z)

BALLOU page 667 Standardised Normal/page 668 Unit Normal Loss Integral
Silver and Peterson page 699 Gu(k) represents the Unit Normal Loss Integral (see also P 271)
HANDBOOK of MATHEMATICAL FUNCTIONS M. Abramowitz and I. A. Stegun
page 966-972, Page 976 table 26.5 , Page 3 pi = 3.1415 92653 58979
FACTOR = Z
P(Z) = (1/sqrt(2*pi))*exp(-1/2*power(Z,2))
AREA = integral -00 to Z over P(Z)
NORM_FUNCTION = exp(-0.5*power(NORM_FACTOR,2))/sqrt(2*3.141592654)

MAU01 /4 - Weighting Schemes

DESCRIPTION

This screen displays graphical and numerical results of the various weighting schemes, one may choose a weighting scheme and calculate weights changing the weighting factor over its range of allowed values.

WEIGHTING SCHEMES

Weighting Scheme W, Weighting Factor F

Unit: W = 1, No weighting
Linear: W = greatest ((Factor-Y), 0)
Reciprocal: W = 1/(1+greatest ((Factor xY), 0))
Reciprocal: W = exp (- Factor xY) AIS Webmaster