SIMULATION MODELING AND ANALYSIS WITH ARENA

T. Altiok and B. Melamed

Chapter 5 Arena Basics

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 1

The Arena Simulation System

Arena is a powerful simulation environment
consists of modeling object templates, called modules, and transactions that move among them, called entities has a visual front-end built around SIMAN block-oriented language constructs and other facilities

SIMAN consists of two classes of objects:

Blocks are basic logic constructs that represent operations, such as SEIZE blocks that model seizing of a facility by a transaction entity, while RELEASE blocks release the facility for use by other transaction entities Elements are objects that represent facilities, such as RESOURCES and QUEUES

Arena modules are are selected from template panels

examples: Basic Process, Advanced Process, Advanced Transfer

Arena modules are high-level constructs that functionally equivalent to sets of SIMAN blocks and/or elements, and internally are built of SIMAN blocks and/or elements
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 2

The Arena Home Screen

Title Bar Menu Bar Run Interaction Toolbar View Toolbar

Standard Toolbar

Drawing Toolbar

Animate Toolbar Animate Transfer Toolbar

Template Panel Project Bar Toolbar Model Window Canvas Flowchart View

Model Window Canvas Spreadsheet View

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 3

Example: A Simple Workstation

Consider a single workstation, known in queueing theory as the M/M/1 queue, where
there is a machine with an infinite buffer in front of it jobs arrive randomly and wait in the buffer while the machine is busy jobs are processed by the machine and then leave the system job inter-arrival times are exponentially distributed with mean 30 minutes job processing times are exponentially distributed with mean 24 minutes

Create 1

Process 1

Dispose 1

0 0

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

Simulation Objects and Actions

Simulating the above workstation calls for the following actions:
jobs are created, one at a time, according to their arrival distribution if the machine is busy processing another job, then the arriving job is queued in the buffer when a job advances to the head of the buffer, it seizes the machine for processing once it becomes available, and holds it for a time period, sampled from its processing-time distribution on process completion, the job departs the machine and is removed from the system

Simulation objects and their actions and interactions

are modeled by Arena modules parameterized by associated dialog boxes

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

Create Module
The Create module generates a stream of arrivals of Arena entities (jobs, people, messages, etc.)

Dialog box for a Create module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 6

Create Module (Cont.)

The Type pull-down menu for the Time Between Arrivals field offers the following options:
Random (exponential inter-arrival times with mean given in the Value field) Schedule (allows the user to create arrival schedules using the Schedule module from the Basic Process template panel Constant (specifies fixed inter arrival times) Expression (any type of inter-arrival time pattern specified by an Arena expression, including Arena distributions)

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

Process Module
The Process module processes (serves) Arena entities

Dialog box for a Process module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 8

Process Module (Cont.)

The Action field option, selected from the pull-down menu, is Seize Delay Release, which stands for a sequence of SEIZE, DELAY and RELEASE SIMAN blocks
SEIZE and RELEASE blocks are used to model contention for a resource possessing a capacity (e.g., machines) when resource capacity is exhausted, the entities contending for the resource must wait until the resource is released thus, the SEIZE block operates like a gate between entities and a resource the processing (holding) time of a resource (called Machine in the example) by an entity is specified via a DELAY block within the Process module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

Dispose Module
The Dispose module implements an entity sunset mechanism
entities that enter it are simply discarded

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

10

Arena Files
Whenever an Arena model is saved, the model is placed in a file with a .doe extension (e.g., mymodel.doe) whenever an Arena model (say, mymodel.doe) is checked using the Check Model option in the Run Menu or any run option in it, Arena automatically creates a number of files:
mymodel.p (program file) mymodel.mdb (Access database file) mymodel.err (errors file) mymodel.opw (model components file) mymodel.out (SIMAN output report file)

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

11

Arena Simulation Results

The end-result of a simulation run is a set of requisite statistics, referred to as run results, such as
mean waiting times buffer content probabilities resource utilization

Arena provides a considerable number of default statistics in a report that is automatically generated at the end of a simulation run in Arena reports Additional statistics can be obtained by adding statistics collection modules in the model, such as
Record (Basic Process template panel) Statistic (Advanced Process template panel)

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

12

Example: a Resources Report

Resources statistics from a single replication of the simple workstation model

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 13

Example: a Queues Report

Queues statistics from a single replication of the simple workstation model

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 14

Arena Data Storage Objects

Arena variables are user-defined global data storage objects used to store and modify state information either at run initialization, or in the course of a run
such (global) variables are visible everywhere in the model, namely, they can be accessed, examined and modified from every component of the model in an Arena program, variables are typically examined in Decide modules and modified in Assign modules

Arena Attributes are data storage objects associated with entities

unlike variables, attributes are local to entities in the sense that each instance of an entity has its own copy of the attributes

Arena expressions can be viewed as specialized variables that store the value of an associated formula (expression)

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

15

Arena Statistics Collection

Statistics collection via the Statistic module
Time-Persistent statistics are time-average statistics (e.g., average queue lengths, server utilization and various probabilities), and this option can be used to estimate any user-defined probability or time average Tally statistics are customer averages, and have to be specified in a Record module to initiate statistics collection, but it is advisable to include their definition in the Statistic module as well, so the entire set of statistics may be viewed in the same spreadsheet for modeling convenience Counter statistics are used to keep track of counts, and have to be specified in a Record module to initiate statistics collection Output statistics are obtained by evaluating an expression at the end of a simulation run, using Arena variables, such as DAVG(S) (the time average of the time-persistent statistic S), TAVG(S) (the average of tally element S), TFIN (simulation completion time), etc. Frequency statistics are used to produce frequency distributions of (random) expressions, such as Arena variables or resource states, allowing the estimation of steady-state probabilities Statistics collection via the Record module is achieved by its proper

placement in the model

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 16

Record Module
The Record module is used by entities to collect statistics at selected locations in the model
the dialog box below the list of statistics types in a Record module

Dialog box for a Record module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 17

Record Module (Cont.)

The Record module has the following options in its Type field:
the Count option maintains a count with a prescribed increment (any real value, positive or negative), which may be defined as any expression or function, and the counter is incremented whenever an entity enters the Record module the Entity Statistics option provides information on entities, such as time and costing/duration information the Time Interval option tallies the difference between the current time and the time stored in a prescribed attribute of the entering entity the Time Between option tallies the time interval between consecutive entries of entities in the Record module (these intervals correspond to inter-departure times from the module, and the reciprocal of the mean inter-departure times is the modules throughput) the Expression option tallies an expression, whose value is recomputed whenever an entity enters the Record module.
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 18

Arena Output Reports

Arena output reports consist of two types: automatic and user-specified An automatic report of summary statistics is generated automatically at the end of a simulation run by a number of Arena constructs, such as entities, queues and resources
those statistics are implicitly specified by the modeler simply by dragging and dropping those modules into an Arena model no further action is required of the user

A user-specified report provides additional statistics via the Statistic module (Advanced Process template panel) and the Record module (Basic Process template panel) obtained by explicitly specifying statistics collection in those modules,
where the Statistic module is specified in a spreadsheet view, and the Record module must be placed in the appropriate location in the model
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 19

Example: Two Processes in Series

Consider a manufacturing network of two workstations in series, consisting of an assembly workstation followed by a painting workstation, where
jobs arrive at the assembly station with exponentially distributed inter-arrival times of mean 5 hours the assembly process always has all the raw materials necessary to carry out the assembly operation the assembly time is uniformly distributed between 2 and 6 hours after the process is completed, a quality control test is performed, and past data reveal that 15% of the jobs fail the test and go back to the assembly operation for rework jobs that pass the test proceed to the painting operation that takes 3 hours for each unit

We are interested in
simulating the system for 100,000 hours estimating process utilizations, average job waiting times and average job flow times (the elapsed time for a job from start to finish).
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 20

The Arena Model

Arena manufacturing model of assembly and painting processes in series

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 21

Assign Module
The Assign module is used by entities to assign values to attributes
the dialog box below assigns an arrival time to a job attribute

Dialog boxes for an Assign module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 22

Decide Module
The Decide module is used by entities to make branching decisions, based on chance or the truth/falsity of prescribed conditions
the dialog box below makes a two-way probabilistic branching decision

Dialog box for a Decide module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 23

Record Module
The Record module is used by entities to collect statistics
the dialog box below tallies job flow times

Dialog box for a Record module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

24

Resources Report

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

25

Queues Report

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

26

User Specified Report

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

27

Example: Production/Inventory System

Consider a production/inventory system in which
the production process (e.g., packaging) is comprised of three stages 1. filling each container unit (e.g., bottles) 2. sealing each unit 3. placing labels on each unit a raw-material storage feeds the production process, and finished units are stored in a warehouse customers arrive at the warehouse with product requests (demands), and if a request cannot be fully satisfied by on-hand inventory, the unsatisfied portion represents lost business
Process Raw Material Lost Customers
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 28

Warehouse Demand

The Production Operation

The production process component operates as follows:
there is always sufficient raw material in storage, so the production process never starves product processing is carried out in lots of 5 units, and finished lots are placed in the warehouse lot processing time is uniformly distributed between 10 and 20 minutes the production process experiences random failures, which may occur at any point in time times between failures are exponentially distributed with a mean of 200 minutes repair times are normally distributed, with a mean of 90 minutes and a standard deviation of 45 minutes

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

29

The Warehouse Operation

The warehouse component operates as follows:
warehouse operations implement the (r, R) inventory control policy, where 1. the warehouse has a capacity (target level) of R = 500 units 2. the production process stops when the inventory in the warehouse reaches the target level 3. from this point and on, the production process remains inactive until the inventory level drops to or below the reorder point of r =150 units 4. At this point the production process is restarted until the warehouse inventory level reaches the target level R = 500 the inter-arrival times between successive customers are uniformly distributed between 3 to 7 hours, and individual demand sizes are distributed uniformly between 50 and 100 units on customer arrival, the inventory is immediately checked, and if there is sufficient stock on hand, that demand is promptly satisfied, and otherwise, the unsatisfied portion of the demand is lost

. .

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

30

System Performance Measures

The performance measures of interest are:
process utilization process downtime probability warehouse average inventory level percentage of customers whose demand is not completely satisfied on arrival at the warehouse average size of lost customer demands at the warehouse, given that the demands are not completely satisfied

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

31

Production/Inventory Arena Model

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

32

Arena Model Logic

The Arena model of the production/inventory system is composed of two segments: inventory management and demand management In the Arena model logic of the inventory management segment:
the packaging process takes a unit of raw material from its queue, and processes it as a batch of 5 the finished lot is added to the warehouse inventory (variable Inventory) if Inventory reaches or up-crosses the target level (variable Target Stock), then production stops until the reorder point (variable Reorder Point) is reached or down-crossed again processing of a new batch starts immediately when the reorder point is down-crossed

In the Arena model logic of the demand management segment:

arrivals of customers and their demands at the warehouse are generated the variable Inventory is adjusted upon customer arrival the value of Inventory is monitored for triggering resumption of suspended production when the reorder point is down-crossed track is kept of lost demand
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 33

Inventory Management Logic

The Arena model logic implementation of the inventory management segment makes use of the following modules:
the Create module, called Raw Material, generates product units for the packaging (batching) operation the packaging operation is modeled using a sequence of Seize, Delay and Release modules the actual processing (packaging in our case) takes place at the Delay module, called Packaging Process, where the packaging time of a batch is specified as Unif(10,20) minutes

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

34

Seize and Release Modules

The Seize module is used by entities to seize resources The Release module is used by entities to release resources
the dialog boxes below seize and release a packaging machine

Dialog boxes for a Seize module and a Release module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 35

Resource Module
A spreadsheet view of Resource modules (from the Basic Process template panel) is used to list all resources and their attributes One of the attribute fields is Failures, which specifies resource failures in a dialog box
the dialog box below shows the resource Packaging Process with its Failures field details

Spreadsheet views for a Resource module (bottom) and its Failures dialog box (top)
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 36

Failure Module
All failure/repair data (uptimes and downtimes) are specified in the spreadsheet Failure module (from the Advanced Process template panel)
the dialog box below illustrates the data for a random failure

Dialog box of a Failure module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

37

Hold Module
The Hold module performs a gating function on entities
the dialog box below illustrates the use of a Hold module for checking the production state

Dialog Box of a Hold module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 38

Hold Module (Cont.)

The state of production (0 = off or 1 = on) is maintained in a variable, called Production, initially set to 0
the Hold module scan for the truth or falsehood of a logical condition, (Production = = 1), and holds the entity in the module until the condition becomes true once the condition becomes true, the entity proceeds to the next module, and otherwise, it waits on in a Queue, called Shall We Produce?.Queue

The inventory level at the warehouse is maintained in a variable, called Inventory, initially set to 250
each product entity (batch of 5 units) that enters the Assign module, called Update Inventory, increases the warehouse inventory level by a batch of 5 finished jobs by adding the Batch Size to the variable Inventory that entity then proceeds from the Hold module, called Shall We Produce?, to the Decide module, called Check Target

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

39

Decide Module
The Decide module is used by entities to make branching decisions, based on chance or the truth/falsity of prescribed conditions
the dialog box below makes a two-way probabilistic branching decision by checking the inventory target level

Dialog Box of a Decide module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 40

Decide Module (Cont.)

The entity tests there whether or not the inventory target level has been up-crossed as follows:
if the inventory target level has been up-crossed, then the entity moves on to the Assign module, called Stop Production, and sets Production = 0 to signal that production is suspended otherwise, the entity does nothing

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

41

Wrapping Up Entity Sojourns

An entity has completed its sojourn through the system, and would normally be disposed of (at a Dispose module) However, since the Packaging module is never starved, and there are no delays incurred since its departure from the Packaging station, we can recycle the entity by always sending it back to the Packaging Queue to play the role of a new arrival This modeling device is logically equivalent to disposing of the entity and creating a new
however, it is computationally more efficient, since it saves us this extra computational effort thus, the simulation will run faster!

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

42

Demand Management Logic

The Arena model logic implementation of the demand management segment makes use of the following modules:
the Create module, called Customer Arrives, generates arrivals of customers, and the Assign module, called Customer Demand, assigns to the customer a demand level the (r,R) inventory policy is modeled using a sequence of Decide and Assign modules the Inventory variable keeps track of the inventory level in the warehouse

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

43

Create Module
The Create module generates a stream of arrivals of Arena entities (jobs, people, demands, etc.)
the dialog box below models the arrival of customer demands

Dialog Box of a Create module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 44

Create and Assign Modules

The arrival pattern of customers is specified to be random with inter-arrival time distribution Unif(3,7) On arrival, the customer entity first enters the Assign module, called Customer Demand, where its Demand attribute is assigned a random value from the distribution Unif(50,100)

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

45

Decide Module
The customer then proceeds to the Decide module, called Check Inventory, to test whether the warehouse has sufficient inventory on hand to satisfy its demand If the variable Inventory has an equal or larger value than attribute Demand, then
the customer takes the True exit to the Assign module, called Take Away From Inventory, where its inventory is decreased the by the demand amount it next proceeds to the Decide module, named Restart Production, to test whether the Reorder Level variable has just been down crossed if it was down crossed, the customer proceeds to the Assign module, called Production Start, to set Production = 1, which would promptly release the product entity currently detained in the Hold module Shall We Produce?, effectively resuming the production process either way, the Customer entity proceeds to be disposed of at the Dispose module, called Dispose1
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 46

Decide Module (Cont.)

If the value of variable Inventory is smaller than attribute Demand, then the current demand is either partially satisfied or not at all
either way, the customer entity proceeds to the Assign module, called Lost Customer, where it sets the Inventory variable to 0 there it updates the variable Lost, which keeps track of customers whose demand could not be fully satisfied (Lost = Lost + 1), and the variable Lost Amount, which keeps track of the customers demand lost (Lost Amount = Demand Inventory) the customer entity next enters the Record module, called Tally Amount Lost, to tally the lost quantity per customer whose demand was not fully satisfied the customer entity then proceeds to be disposed of at the module called Dispose1

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

47

Variable Module
The Variable spreadsheet module (Basic Process template panel) is used to set or inspected user-defined variables defined in the model and their properties
the dialog box bellow lists user defined variables for our example system

Dialog Box of a Variable module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 48

Statistic Module
The Statistic spreadsheet module (Advanced Process template panel) is used to specify statistics collection in the model and their properties
the dialog box bellow lists statistics collected in our example system

Dialog Box of a Statistic module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

49

Statistic Module (Cont.)

Statistics listed include
a Time-Persistent statistic, called Stock On Hand, for the Inventory variable a Time-Persistent statistic, called Production On, for the condition Production = 1 a Frequency statistic, called Process States, which estimates the state probabilities of the packaging process, namely, the probabilities that the packaging process is busy or down

The statistical outputs for Inventory include

the average value 95% confidence interval minimal and maximal values The statistical output for state Production = 1 is the percentage of

time this expression is true, that is, the probability that the packaging process is in production
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 50

Record Module
The Record module is used to specify user-defined statistics collection via entities in particular model locations
the dialog box bellow specifies the tallying of demand loss at a Record module, called Tally Lost Amount

Dialog Box of a Record module

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 51

Record Module (Cont.)

Whenever a customer entity enters the module, the expression (in this case, the variable) Amount Lost is evaluated, and the resultant value is tallied When the replication terminates, the output report will contain a Tallies section summarizing the statistics of the amount lost per customer

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

52

Simulation Output Reports

The simulation was run for one replication of length 1,000,000 minutes (slightly less than two years) The reports produced include User Specified report Frequencies report

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

53

User Specified Report

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

54

Frequencies Report

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

55

Experimentation and Analysis

Clearly, the customer service level (probability that the demand of an arriving customer is fully satisfied) is quite low as measured by the complementary probability of partially or fully unsatisfied demands, which was estimated by the simulation as 0.25 (the value of Lost Percentage in the Output section of the User Specified report) We want to modify the system to increase the customer service level (equivalently, we want to decrease the loss probability from 0.25 to an acceptable level)
in inventory-oriented systems, such as the one under consideration, the only way to increase the customer service level is to increase the level of inventory on-hand

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

56

Improvement Strategy 1
In our case, we may attempt to achieve the goal of increasing the customer service level is to use Strategy 1 for modifying the original Production/Inventory system as follows:
invest more in maintenance activities this would reduce downtimes and consequently make the process more available for production

The next two reports show the improvements under this strategy (Strategy 1)
the average repair time is reduced from 90 minutes to 70 minutes consequently, the loss probability is reduced from 0.25 to 0.08!

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

57

Frequencies Report for Strategy 1

The Frequencies report below indicates that Strategy 1 reduces the average repair time to 70 minutes with a standard deviation of 25 minutes

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

58

User Specified Report for Strategy 1

The User Specified report below indicates that Strategy 1 reduces the loss probability to 0.08

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

59

Improvement Strategy 2
Strategy 2 increases the reorder level so that the reorder level is hit sooner and production would resume earlier
when the reorder level is doubled, then the amount lost per customers is marginally improved from 25% to about 21.5% however, we would like to do better as in Strategy 1

This analysis leads us to conclude the following:

a significant improvement in the customer service level can be achieved by improving the production process, rather than modifying the inventory replenishment policy

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5

60

Time-Dependent Arrivals
So far we assumed that random phenomena (e.g., arrivals, services, etc.) are modeled as variates from a fixed probability law that does not change in time, so that the underlying processes are stationary However, it is quite common in practice for the underlying probability law to vary in time, in which case the underlying process is non-stationary (time dependent)
for example, rush hour traffic as opposed to ebb hour traffic

The Arena Create module provides a Schedule option in its Time Between Arrivals section, to be used only for single parameter distributions
example: the exponential distribution, which gives rise to the Poisson arrival process
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 61

Schedule Option of Create Module

Dialog box of a Create module (bottom) and spreadsheet view of its Schedule option (top)
Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 62

Schedule Option Specification

Dialog box for a schedule specifying time-dependent arrival rates

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 63

Schedule Option Specification (Cont.)

Dialog box for the Options field of a schedule

Altiok / Melamed Simulation Modeling and Analysis with Arena Chapter 5 64