Package 'simmer'

simmer: Discrete-Event Simulation for R

Description

A process-oriented and trajectory-based Discrete-Event Simulation (DES) package for R. Designed to be a generic framework like SimPy or SimJulia, it leverages the power of Rcpp to boost the performance and turning DES in R feasible. As a noteworthy characteristic, simmer exploits the concept of trajectory: a common path in the simulation model for entities of the same type. It is pretty flexible and simple to use, and leverages the chaining/piping workflow introduced by the magrittr package.

Author(s)

Iñaki Ucar and Bart Smeets

References

Ucar I., Smeets B., Azcorra A. (2019). "simmer: Discrete-Event Simulation for R." Journal of Statistical Software, 90(2), 1-30. doi:10.18637/jss.v090.i02.

Ucar I., Hernández J.A., Serrano P., Azcorra A. (2018). "Design and Analysis of 5G Scenarios with simmer: An R Package for Fast DES Prototyping." IEEE Communications Magazine, 56(11), 145-151. doi:10.1109/MCOM.2018.1700960.

See Also

simmer's homepage https://r-simmer.org and GitHub repository https://github.com/r-simmer/simmer.

Examples

## Not run: 
# introduction to simmer
vignette("simmer-01-introduction")

# JSS paper available as vignette
vignette("simmer-02-jss")

# more vignettes
vignette(package = "simmer")

## End(Not run)

---

Activate/Deactivate Sources

Description

Activities for activating or deactivating the generation of arrivals by name. Sources must be defined in the simulation environment (see add_generator, add_dataframe).

Usage

activate(.trj, sources, ..., tag)

deactivate(.trj, sources, ..., tag)

Arguments

.trj	the trajectory object.
sources	the name(s) of the source(s) or a function returning the name(s).
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.

Value

Returns the trajectory object.

See Also

set_trajectory, set_source

Examples

traj <- trajectory() %>%
  deactivate("dummy") %>%
  timeout(1) %>%
  activate("dummy")

simmer() %>%
  add_generator("dummy", traj, function() 1) %>%
  run(10) %>%
  get_mon_arrivals()

---

Add a Data Frame

Description

Attach a new source of arrivals to a trajectory from a data frame.

Usage

add_dataframe(.env, name_prefix, trajectory, data, mon = 1, batch = 50,
  col_time = "time", time = c("interarrival", "absolute"),
  col_attributes = NULL, col_priority = "priority",
  col_preemptible = col_priority, col_restart = "restart")

Arguments

.env	the simulation environment.
name_prefix	the name prefix of the generated arrivals.
trajectory	the trajectory that the generated arrivals will follow (see trajectory).
data	a data frame with, at least, a column of (inter)arrival times (see details).
mon	whether the simulator must monitor the generated arrivals or not (0 = no monitoring, 1 = simple arrival monitoring, 2 = level 1 + arrival attribute monitoring)
batch	number of arrivals generated at a time. Arrivals are read from the data frame and attached to the trajectory in batches depending on this value. In general, it should not be changed.
col_time	name of the time column in the data frame.
time	type of time column: interarrival, if the time column contains interarrival times, or absolute, if the time column contains absolute arrival times.
col_attributes	vector of names of the attributes columns (see details).
col_priority	name of the priority column.
col_preemptible	name of the preemptible column.
col_restart	name of the restart column.

Details

The data frame provided must have, at least, a column of (inter)arrival times. This method will look for it under the name "time" by default, although this can be changed with the col_time parameter.

If there is any column named col_priority="priority", col_preemptible=priority or col_restart="restart", they will be used to set the prioritization values for each arrival (see add_generator).

If there are additional columns (with col_attributes=NULL, by default), they will be assigned to arrival attributes named after each column name. All these columns must be numeric (or logical). Otherwise, if a vector of column names is specified, only these will be assigned as attributes and the rest of the columns will be ignored.

A value of batch=Inf means that the whole data frame will be attached at the beginning of the simulation. This is not desirable in general, because the performance of the event queue is degraded when it is populated with too many events. On the other hand, a low value results in an increased overhead due to many function calls. The default value has been tested to provide a good trade-off.

Value

Returns the simulation environment.

See Also

Other sources: add_generator.

---

Add a Generator

Description

Attach a new source of arrivals to a trajectory from a generator function.

Usage

add_generator(.env, name_prefix, trajectory, distribution, mon = 1,
  priority = 0, preemptible = priority, restart = FALSE)

Arguments

.env	the simulation environment.
name_prefix	the name prefix of the generated arrivals. If several names are provided, several generators will be defined with the same parameters.
trajectory	the trajectory that the generated arrivals will follow (see trajectory).
distribution	a function modelling the interarrival times (returning a negative value or a missing value stops the generator).
mon	whether the simulator must monitor the generated arrivals or not (0 = no monitoring, 1 = simple arrival monitoring, 2 = level 1 + arrival attribute monitoring)
priority	the priority of each arrival (a higher integer equals higher priority; defaults to the minimum priority, which is 0).
preemptible	if a seize occurs in a preemptive resource, this parameter establishes the minimum incoming priority that can preempt these arrivals (an arrival with a priority greater than preemptible gains the resource). In any case, preemptible must be equal or greater than priority, and thus only higher priority arrivals can trigger preemption.
restart	whether the activity must be restarted after being preempted.

Value

Returns the simulation environment.

See Also

Convenience functions: at, from, to, from_to, when_activated.

Other sources: add_dataframe.

---

Add a Global Attribute

Description

Attach a global variable to the simulation.

Usage

add_global(.env, key, value)

Arguments

.env	the simulation environment.
key	the attribute name.
value	the value to set, either a numeric or a schedule, so that the global may change during the simulation.

Value

Returns the simulation environment.

See Also

Convenience functions: schedule.

---

Add a Resource

Description

Define a new resource in a simulation environment. Resources are conceived with queuing systems in mind, and therefore they comprise two internal self-managed parts: a server, which is the active part, with a specified capacity that can be seized and released (see seize); and a priority queue of a certain size, in which arrivals may wait for the server to be available.

Usage

add_resource(.env, name, capacity = 1, queue_size = Inf, mon = TRUE,
  preemptive = FALSE, preempt_order = c("fifo", "lifo"),
  queue_size_strict = FALSE, queue_priority = c(0, Inf))

Arguments

.env	the simulation environment.
name	the name of the resource. If several names are provided, several resources will be defined with the same parameters.
capacity	the capacity of the server, either an integer or a schedule, so that the value may change during the simulation.
queue_size	the size of the queue, either an integer or a schedule, so that the value may change during the simulation.
mon	whether the simulator must monitor this resource or not.
preemptive	whether arrivals in the server can be preempted or not based on seize priorities.
preempt_order	if preemptive=TRUE and several arrivals are preempted, this parameter defines which arrival should be preempted first. Either fifo (First In First Out: older preemptible tasks are preempted first) or lifo (Last In First Out: newer preemptible tasks are preempted first).
queue_size_strict	whether the queue_size is a hard limit (see details).
queue_priority	the priority range required to be able to access the queue if there is no room in the server (if a single value is provided, it is treated as the minimum priority). By default, all arrivals can be enqueued.

Details

An entity trying to seize a resource (see seize) may 1) access the server straightaway if there is enough capacity, 2) wait in the queue if there is no room in the server but there is room in the queue, or 3) rejected if there is no room in the queue either.

There are two special situations regarding queue management: 1) the queue_size is shrinked below the actual number of items waiting, and 2) preemption occurs, and an item previously in the server goes to the queue. By default in both cases, the excess of items in the queue is allowed. However, with queue_size_strict=TRUE, the maximum queue_size is guaranteed, and thus some entities will be rejected (dropped) by the resource.

Whenever an arrival is rejected (due to a server drop or a queue drop), it will set the finished flag to FALSE in the output of get_mon_arrivals. Unfinished arrivals can be handled with a drop-out trajectory that can be set using the handle_unfinished activity.

Value

Returns the simulation environment.

See Also

Convenience functions: schedule.

---

Batch/Separate Arrivals

Description

Activities for collecting a number of arrivals before they can continue processing and splitting a previously established batch.

Usage

batch(.trj, n, timeout = 0, permanent = FALSE, name = "", rule = NULL,
  ..., tag)

separate(.trj, ..., tag)

Arguments

.trj	the trajectory object.
n	batch size, accepts a numeric or a callable object (a function) which must return a numeric.
timeout	set an optional timer which triggers batches every timeout time units even if the batch size has not been fulfilled, accepts a numeric or a callable object (a function) which must return a numeric (0 = disabled).
permanent	if TRUE, batches cannot be split.
name	optional string. Unnamed batches from different batch activities are independent. However, if you want to feed arrivals from different trajectories into a same batch, you need to specify a common name across all your batch activities.
rule	an optional callable object (a function) which will be applied to every arrival to determine whether it should be included into the batch, thus it must return a boolean.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.

Value

Returns the trajectory object.

Examples

## unnamed batch with a timeout
traj <- trajectory() %>%
  log_("arrived") %>%
  batch(2, timeout=5) %>%
  log_("in a batch") %>%
  timeout(5) %>%
  separate() %>%
  log_("leaving")

simmer() %>%
  add_generator("dummy", traj, at(0:2)) %>%
  run() %>% invisible

## batching based on some dynamic rule
traj <- trajectory() %>%
  log_("arrived") %>%
  # always FALSE -> no batches
  batch(2, rule=function() FALSE) %>%
  log_("not in a batch") %>%
  timeout(5) %>%
  separate() %>%
  log_("leaving")

simmer() %>%
  add_generator("dummy", traj, at(0:2)) %>%
  run() %>% invisible

## named batch, shared across trajectories
traj0 <- trajectory() %>%
  log_("arrived traj0") %>%
  batch(2, name = "mybatch")

traj1 <- trajectory() %>%
  log_("arrived traj1") %>%
  timeout(1) %>%
  batch(2, name = "mybatch") %>%
  log_("in a batch") %>%
  timeout(2) %>%
  separate() %>%
  log_("leaving traj1")

simmer() %>%
  add_generator("dummy0", traj0, at(0)) %>%
  add_generator("dummy1", traj1, at(0)) %>%
  run() %>% invisible

---

Fork the Trajectory Path

Description

Activity for defining a fork with N alternative sub-trajectories.

Usage

branch(.trj, option, continue, ..., tag)

Arguments

.trj	the trajectory object.
option	a callable object (a function) which must return an integer between 0 and N. A return value equal to 0 skips the branch and continues to the next activity. A returning value between 1 to N makes the arrival to follow the corresponding sub-trajectory.
continue	a vector of N booleans that indicate whether the arrival must continue to the main trajectory after each sub-trajectory or not (if only one value is provided, it will be recycled to match the number of sub-trajectories).
...	N trajectory objects (or a list of N trajectory objects) describing each sub-trajectory.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.

Value

Returns the trajectory object.

Examples

env <- simmer()

traj <- trajectory() %>%
  set_global("path", 1, mod="+", init=-1) %>%
  log_(function() paste("Path", get_global(env, "path"), "selected")) %>%
  branch(
    function() get_global(env, "path"), continue=c(TRUE, FALSE),
    trajectory() %>%
      log_("following path 1"),
    trajectory() %>%
      log_("following path 2")) %>%
  log_("continuing after the branch (path 0)")

env %>%
  add_generator("dummy", traj, at(0:2)) %>%
  run() %>% invisible

---

Clone/Synchronize Arrivals

Description

Activities for defining a parallel fork and removing the copies. clone replicates an arrival n times (the original one + n-1 copies). synchronize removes all but one clone for each set of clones.

Usage

clone(.trj, n, ..., tag)

synchronize(.trj, wait = TRUE, mon_all = FALSE, ..., tag)

Arguments

.trj	the trajectory object.
n	number of clones, accepts either a numeric or a callable object (a function) which must return a numeric.
...	a number of optional parallel sub-trajectories (or a list of sub-trajectories). Each clone will follow a different sub-trajectory if available.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.
wait	if FALSE, all clones but the first to arrive are removed. if TRUE (default), all clones but the last to arrive are removed.
mon_all	if TRUE, get_mon_arrivals will show one line per clone.

Value

Returns the trajectory object.

Examples

## clone and wait for the others
traj <- trajectory() %>%
  clone(
    n = 3,
    trajectory() %>%
      log_("clone 0 (original)") %>%
      timeout(1),
    trajectory() %>%
      log_("clone 1") %>%
      timeout(2),
    trajectory() %>%
      log_("clone 2") %>%
      timeout(3)) %>%
  log_("sync reached") %>%
  synchronize(wait = TRUE) %>%
  log_("leaving")

simmer() %>%
  add_generator("arrival", traj, at(0)) %>%
  run() %>% invisible

## more clones that trajectories available
traj <- trajectory() %>%
  clone(
    n = 5,
    trajectory() %>%
      log_("clone 0 (original)") %>%
      timeout(1)) %>%
  log_("sync reached") %>%
  synchronize(wait = TRUE) %>%
  log_("leaving")

simmer() %>%
  add_generator("arrival", traj, at(0)) %>%
  run() %>% invisible

## clone and continue without waiting
traj <- trajectory() %>%
  clone(
    n = 3,
    trajectory() %>%
      log_("clone 0 (original)") %>%
      timeout(1),
    trajectory() %>%
      log_("clone 1") %>%
      timeout(2),
    trajectory() %>%
      log_("clone 2") %>%
      timeout(3)) %>%
  log_("sync reached") %>%
  synchronize(wait = FALSE) %>%
  log_("leaving")

simmer() %>%
  add_generator("arrival", traj, at(0)) %>%
  run() %>% invisible

---

Extract or Replace Parts of a Trajectory

Description

Operators acting on trajectories to extract or replace parts.

Usage

## S3 method for class 'trajectory'
x[i]

## S3 method for class 'trajectory'
x[[i]]

## S3 replacement method for class 'trajectory'
x[i] <- value

## S3 replacement method for class 'trajectory'
x[[i]] <- value

Arguments

x	the trajectory object.
i	indices specifying elements to extract. Indices are numeric or character or logical vectors or empty (missing) or NULL. Numeric values are coerced to integer as by as.integer (and hence truncated towards zero). Negative integers indicate elements/slices to leave out the selection. Character vectors will be matched to the names and tags of the activities in the trajectory as by %in%. Logical vectors indicate elements/slices to select. Such vectors are recycled if necessary to match the corresponding extent. An empty index will return the whole trajectory. An index value of NULL is treated as if it were integer(0).
value	another trajectory object.

Value

Returns a new trajectory object.

See Also

length.trajectory, get_n_activities, join.

Examples

x <- join(lapply(1:12, function(i)
  trajectory() %>% timeout(i)
))
x

x[10]                 # the tenth element of x
x[-1]                 # delete the 1st element of x
x[c(TRUE, FALSE)]     # logical indexing
x[c(1, 5, 2, 12, 4)]  # numeric indexing
x[c(FALSE, TRUE)] <- x[c(TRUE, FALSE)] # replacing
x

---

Convenience Functions for Generators

Description

These convenience functions facilitate the definition of generators of arrivals for some common cases.

Usage

at(...)

from(start_time, dist, arrive = TRUE)

to(stop_time, dist)

from_to(start_time, stop_time, dist, arrive = TRUE, every = NULL)

when_activated(n = 1)

Arguments

...	a vector or multiple parameters of times at which to initiate an arrival.
start_time	the time at which to launch the initial arrival (numeric or function).
dist	a function modelling the interarrival times. It is supposed to be an infinite source of values >= 0 (e.g., rexp and the like). If the function provided returns any negative value, the behaviour is undefined.
arrive	if set to TRUE (default) the first arrival will be generated at start_time and will follow dist from then on. If set to FALSE, will initiate dist at start_time (and the first arrival will most likely start at a time later than start_time).
stop_time	the time at which to stop the generator (numeric or function).
every	repeat with this time cycle (numeric or function).
n	an integer or a callable object (a function) which must return a number of arrivals to generate when activated.

Details

at generates arrivals at specific absolute times.

from generates inter-arrivals following a given distribution with a specified start time. union of the last two.

to generates inter-arrivals following a given distribution with a specified stop time.

from_to is the union of from and to.

when_activated sets up an initially inactive generator which generates n arrivals each time it is activated from any trajectory using the activity activate.

Value

Returns a generator function (a closure).

See Also

add_generator.

Examples

## common to all examples below
# some trajectory
t0 <- trajectory() %>%
  timeout(0)
# some distribution
distr <- function() runif(1, 1, 2)

# arrivals at 0, 1, 10, 30, 40 and 43
simmer() %>%
  add_generator("dummy", t0, at(0, c(1,10,30), 40, 43)) %>%
  run(100) %>%
  get_mon_arrivals()

# apply distribution starting at 5 (and no end)
simmer() %>%
  add_generator("dummy", t0, from(5, distr)) %>%
  run(10) %>%
  get_mon_arrivals()

# apply distribution until 5 (starting at 0)
simmer() %>%
  add_generator("dummy", t0, to(5, distr)) %>%
  run(10) %>%
  get_mon_arrivals()

# apply distribution from 8 to 16 h every 24 h:
simmer() %>%
  add_generator("dummy", t0, from_to(8, 16, distr, every=24)) %>%
  run(48) %>%
  get_mon_arrivals()

# triggering arrivals on demand from a trajectory
t1 <- trajectory() %>%
  activate("dummy")

simmer() %>%
  add_generator("dummy", t0, when_activated()) %>%
  add_generator("trigger", t1, at(2)) %>%
  run() %>%
  get_mon_arrivals()

---

Get Resource Parameters

Description

Getters for resources: server capacity/count and queue size/count, seized amount, activity time, and selected resources.

Usage

get_capacity(.env, resources)

get_capacity_selected(.env, id = 0)

get_queue_size(.env, resources)

get_queue_size_selected(.env, id = 0)

get_server_count(.env, resources)

get_server_count_selected(.env, id = 0)

get_queue_count(.env, resources)

get_queue_count_selected(.env, id = 0)

get_seized(.env, resources)

get_seized_selected(.env, id = 0)

get_activity_time(.env, resources)

get_activity_time_selected(.env, id = 0)

get_selected(.env, id = 0)

Arguments

.env	the simulation environment.
resources	one or more resource names.
id	selection identifier for nested usage.

Details

If no resources are provided to get_activity_time, the overall activity time is reported.

Value

Return a vector (character for get_selected, numeric for the rest of them).

See Also

get_resources, set_capacity, set_queue_size, seize, timeout.

---

Monitoring Statistics

Description

Getters for obtaining monitored data (if any) about arrivals, attributes and resources.

Usage

get_mon_arrivals(.envs, per_resource = FALSE, ongoing = FALSE)

get_mon_attributes(.envs)

get_mon_resources(.envs)

Arguments

.envs	the simulation environment (or a list of environments).
per_resource	if TRUE, statistics will be reported on a per-resource basis.
ongoing	if TRUE, ongoing arrivals will be reported. The columns end_time and finished of these arrivals are reported as NAs.

Value

Returns a data frame.

---

Get Process Parameters

Description

Getters for processes (sources and arrivals) number of arrivals generated by a source, the name of the active arrival, an attribute from the active arrival or a global one, prioritization values, or the number of arrivals in an active batch.

Usage

get_n_generated(.env, sources)

get_trajectory(.env, sources)

get_name(.env)

get_start_time(.env)

get_attribute(.env, keys)

get_global(.env, keys)

get_prioritization(.env)

get_batch_size(.env)

Arguments

.env	the simulation environment.
sources	one or more resource names.
keys	the attribute name(s).

Details

get_n_generated returns the number of arrivals generated by the given sources. get_trajectory returns the trajectory to which they are attached (as a list).

get_name returns the number of the running arrival. get_start_time returns the start time of the running arrival. get_attribute returns a running arrival's attributes. If a provided key was not previously set, it returns a missing value. get_global returns a global attribute. get_prioritization returns a running arrival's prioritization values. get_name, get_start_time, get_attribute and get_prioritization are meant to be used inside a trajectory; otherwise, there will be no arrival running and these functions will throw an error.

See Also

get_sources, set_trajectory, set_attribute, set_global, set_prioritization, batch.

---

Get Sources and Resources Defined

Description

Get a list of names of sources or resources defined in a simulation environment.

Usage

get_sources(.env)

get_resources(.env)

Arguments

.env	the simulation environment.

Value

A character vector.

---

Handle Unfinished Arrivals

Description

Activity for setting a drop-out trajectory for unfinished arrivals, i.e., those dropped from a resource (due to preemption, resource shrinkage or a rejected seize) or those that leave a trajectory.

Usage

handle_unfinished(.trj, handler, ..., tag)

Arguments

.trj	the trajectory object.
handler	trajectory object to handle unfinished arrivals. A NULL value will unset the drop-out trajectory.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.

Value

Returns the trajectory object.

See Also

leave, set_capacity

Examples

traj <- trajectory() %>%
  log_("arrived") %>%
  handle_unfinished(
    trajectory() %>%
      log_("preempted!")) %>%
  seize("res") %>%
  log_("resource seized") %>%
  timeout(10) %>%
  release("res") %>%
  log_("leaving")

simmer() %>%
  add_resource("res", 1, 0, preemptive=TRUE, queue_size_strict=TRUE) %>%
  add_generator("dummy", traj, at(0)) %>%
  add_generator("priority_dummy", traj, at(5), priority=1) %>%
  run() %>% invisible

---

Join Trajectories

Description

Concatenate any number of trajectories in the specified order.

Usage

join(...)

Arguments

...	trajectory objects.

Value

Returns a new trajectory object.

See Also

Extract.trajectory, length.trajectory, get_n_activities.

Examples

t1 <- trajectory() %>% seize("dummy", 1)
t2 <- trajectory() %>% timeout(1)
t3 <- trajectory() %>% release("dummy", 1)

## join can be used alone
join(t1, t2, t3)

## or can be chained in a trajectory definition
trajectory() %>%
  join(t1) %>%
  timeout(1) %>%
  join(t3)

---

Number of Activities in a Trajectory

Description

Get the number of activities in a trajectory. length returns the number of first-level activities (sub-trajectories not included). get_n_activities returns the total number of activities (sub-trajectories included).

Usage

## S3 method for class 'trajectory'
length(x)

get_n_activities(x)

Arguments

x	the trajectory object.

Value

Returns a non-negative integer of length 1.

See Also

Extract.trajectory, join.

Examples

x <- trajectory() %>%
  timeout(1)

x <- x %>%
  clone(2, x, x)
x

## length does not account for subtrajectories
length(x)
get_n_activities(x)

---

Debugging

Description

Activities for displaying messages preceded by the simulation time and the name of the arrival, and for setting conditional breakpoints.

Usage

log_(.trj, message, level = 0, ..., tag)

stop_if(.trj, condition, ..., tag)

Arguments

.trj	the trajectory object.
message	the message to display, accepts either a string or a callable object (a function) which must return a string.
level	debugging level. The message will be printed if, and only if, the level provided is less or equal to the log_level defined in the simulation environment (see simmer).
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.
condition	a boolean or a function returning a boolean.

Value

Returns the trajectory object.

Examples

## log levels
traj <- trajectory() %>%
  log_("this is always printed") %>% # level = 0 by default
  log_("this is printed if `log_level>=1`", level = 1) %>%
  log_("this is printed if `log_level>=2`", level = 2)

simmer() %>%
  add_generator("dummy", traj, at(0)) %>%
  run() %>% invisible

simmer(log_level = 1) %>%
  add_generator("dummy", traj, at(0)) %>%
  run() %>% invisible

simmer(log_level = Inf) %>%
  add_generator("dummy", traj, at(0)) %>%
  run() %>% invisible

---

Create a Monitor

Description

Methods for creating monitor objects for simulation environments.

Usage

monitor(name, xptr, get_arrivals, get_attributes, get_resources,
  handlers = NULL, finalizer = NULL)

monitor_mem()

monitor_delim(path = tempdir(), keep = FALSE, sep = " ", ext = ".txt",
  reader = read.delim, args = list(stringsAsFactors = FALSE))

monitor_csv(path = tempdir(), keep = FALSE, reader = read.csv,
  args = list(stringsAsFactors = FALSE))

Arguments

name	an identifier to show when printed.
xptr	an external pointer pointing to a C++ object derived from the abstract class simmer::Monitor. See C++ API for further details and, in particular, the simmer/monitor.h header.
get_arrivals	a function to retrieve the arrivals tables. It must accept the xptr as a first argument, even if it is not needed, and a boolean per_resource as a second argument (see get_mon_arrivals).
get_attributes	a function to retrieve the attributes table. It must accept the xptr as a first argument, even if it is not needed.
get_resources	a function to retrieve the resources table. It must accept the xptr as a first argument, even if it is not needed.
handlers	an optional list of handlers that will be stored in a slot of the same name. For example, monitor_mem does not use this slot, but monitor_delim and monitor_csv store the path to the created files.
finalizer	an optional one-argument function to be called when the object is destroyed. For example, monitor_mem does not require any finalizer, but monitor_delim and monitor_csv use this to remove the created files when the monitor is destroyed.
path	directory where files will be created (must exist).
keep	whether to keep files on exit. By default, files are removed.
sep	separator character.
ext	file extension to use.
reader	function that will be used to read the files.
args	a list of further arguments for reader.

Details

The monitor method is a generic function to instantiate a monitor object. It should not be used in general unless you want to extend simmer with a custom monitor.

The in-memory monitor is enabled by default (memory_mem), and it should the fastest.

For large simulations, or if the RAM footprint is an issue, you may consider monitoring to disk. To that end, monitor_delim stores the values in flat delimited files. The usual get_mon_* methods retrieve data frames from such files using the reader provided. By default, read.delim is used, but you may consider using faster alternatives from other packages. It is also possible to keep the files in a custom directory to read and post-process them in a separate workflow.

monitor_csv is a special case of monitor_delim with sep="," and ext=".csv".

Value

A monitor object.

Examples

mon <- monitor_csv()
mon

env <- simmer(mon=mon) %>%
  add_generator("dummy", trajectory() %>% timeout(1), function() 1) %>%
  run(10)
env

read.csv(mon$handlers$arrivals) # direct access
get_mon_arrivals(env)           # adds the "replication" column

---

Simulation Time

Description

Get the current simulation time.

Usage

now(.env)

Arguments

.env	the simulation environment.

Value

Returns a numeric value.

See Also

peek.

---

Peek Next Events

Description

Look for future events in the event queue and (optionally) obtain info about them.

Usage

peek(.env, steps = 1, verbose = FALSE)

Arguments

.env	the simulation environment.
steps	number of steps to peek.
verbose	show additional information (i.e., the name of the process) about future events.

Value

Returns numeric values if verbose=F and a data frame otherwise.

See Also

now.

---

Renege on some Condition

Description

Activities for leaving with some probability, or for setting or unsetting a timer or a signal after which the arrival will abandon.

Usage

leave(.trj, prob, out = NULL, keep_seized = TRUE, ..., tag)

renege_in(.trj, t, out = NULL, keep_seized = FALSE, ..., tag)

renege_if(.trj, signal, out = NULL, keep_seized = FALSE, ..., tag)

renege_abort(.trj, ..., tag)

Arguments

.trj	the trajectory object.
prob	a probability or a function returning a probability.
out	optional sub-trajectory in case of reneging.
keep_seized	whether to keep already seized resources. By default, all resources are released.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.
t	timeout to trigger reneging, accepts either a numeric or a callable object (a function) which must return a numeric.
signal	signal to trigger reneging, accepts either a string or a callable object (a function) which must return a string.

Details

Arrivals that leave the trajectory will set the finished flag to FALSE in the output of get_mon_arrivals. Unfinished arrivals can be handled with a drop-out trajectory that can be set using the optional argument out or the handle_unfinished activity.

Note that, for historical reasons, leave has keep_seized=TRUE by default, while renege_* does not.

Note that renege_if works similarly to trap, but in contrast to that, reneging is triggered even if the arrival is waiting in a queue or is part of a non-permanent batch.

Value

Returns the trajectory object.

See Also

handle_unfinished

send

Examples

## leave with some probability
set.seed(1234)

traj <- trajectory() %>%
  log_("leave with some probability") %>%
  leave(function() runif(1) < 0.5) %>%
  log_("didn't leave")

simmer() %>%
  add_generator("dummy", traj, at(0, 1)) %>%
  run() %>% invisible

## reneging after some time
bank <- trajectory() %>%
  log_("here I am") %>%
  # renege in 5 minutes
  renege_in(
    5,
    out = trajectory() %>%
      log_("lost my patience. Reneging...")) %>%
  seize("clerk") %>%
  # stay if I'm being attended within 5 minutes
  renege_abort() %>%
  log_("I'm being attended") %>%
  timeout(10) %>%
  release("clerk") %>%
  log_("finished")

simmer() %>%
  add_resource("clerk", 1) %>%
  add_generator("customer", bank, at(0, 1)) %>%
  run() %>% invisible

---

Reset a Simulator

Description

Reset the following components of a simulation environment: time, event queue, resources, sources and statistics.

Usage

reset(.env)

Arguments

.env	the simulation environment.

Value

Returns the simulation environment.

See Also

stepn, run.

---

Rollback to a Previous Activity

Description

Activity for going backwards to a previous point in the trajectory. Useful to implement loops.

Usage

rollback(.trj, target, times = Inf, check = NULL, ..., tag)

Arguments

.trj	the trajectory object.
target	tag name (previously set with the tag argument in any activity) or amount of activities (of the same or parent trajectories) to roll back (see examples).
times	the number of repetitions until an arrival may continue.
check	a callable object (a function) which must return a boolean. If present, the times parameter is ignored, and the activity uses this function to check whether the rollback must be done or not.
...	unused
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.

Value

Returns the trajectory object.

Examples

## rollback a specific number of times
traj <- trajectory() %>%
  log_("hello!") %>%
  timeout(1) %>%
  rollback(2, 3)

simmer() %>%
  add_generator("hello_sayer", traj, at(0)) %>%
  run() %>% invisible

## same but with a tag as target
traj <- trajectory() %>%
  log_("hello!", tag="msg") %>%
  timeout(1) %>%
  rollback("msg", 3)

simmer() %>%
  add_generator("hello_sayer", traj, at(0)) %>%
  run() %>% invisible

## custom check
env <- simmer()

traj <- trajectory() %>%
  set_attribute("var", 0) %>%
  log_(tag="msg", function()
    paste("attribute level is at:", get_attribute(env, "var"))) %>%
  set_attribute("var", 25, mod="+") %>%
  rollback("msg", check=function() get_attribute(env, "var") < 100) %>%
  log_("done")

env %>%
  add_generator("dummy", traj, at(0)) %>%
  run() %>% invisible

---

Run a Simulation

Description

Execute steps until a given criterion.

Usage

run(.env, until = Inf, progress = NULL, steps = 10)

stepn(.env, n = 1)

Arguments

.env	the simulation environment.
until	stop time.
progress	optional callback to show the progress of the simulation. The completed ratio is periodically passed as argument to the callback.
steps	number of steps to show as progress (it takes effect only if progress is provided).
n	number of events to simulate.

Value

Returns the simulation environment.

See Also

reset.

Examples

## show the progress just printing the steps
simmer() %>%
  run(progress=message, steps=5)

## using the 'progress' package
## Not run: 
mm1 <- trajectory() %>%
  seize("server", 1) %>%
  timeout(function() rexp(1, 66)) %>%
  release("server", 1)

simmer() %>%
  add_resource("server", 1) %>%
  add_generator("customer", mm1, function() rexp(100, 60)) %>%
  run(3000, progress=progress::progress_bar$new()$update)

## End(Not run)

---

Generate a Scheduling Object

Description

Resource convenience function to generate a scheduling object from a timetable specification.

Usage

schedule(timetable, values, period = Inf)

Arguments

timetable	absolute points in time in which the desired value changes.
values	one value for each point in time.
period	period of repetition.

Value

Returns a schedule object.

See Also

add_resource.

Examples

# Schedule 3 units from 8 to 16 h
#          2 units from 16 to 24 h
#          1 units from 24 to 8 h
capacity_schedule <- schedule(c(8, 16, 24), c(3, 2, 1), period=24)

env <- simmer() %>%
  add_resource("dummy", capacity_schedule)

# Composition of schedules
sch1 <- schedule(c(8, 16), c(3, 0), period=24)
sch2 <- schedule(c(16, 24), c(2, 1), period=24)
all.equal(sch1 + sch2, capacity_schedule)

---

Seize/Release Resources

Description

Activities for seizing/releasing a resource, by name or a previously selected one. Resources must be defined in the simulation environment (see add_resource).

Usage

seize(.trj, resource, amount = 1, continue = NULL, post.seize = NULL,
  reject = NULL, ..., tag)

seize_selected(.trj, amount = 1, id = 0, continue = NULL,
  post.seize = NULL, reject = NULL, ..., tag)

release(.trj, resource, amount = 1, ..., tag)

release_selected(.trj, amount = 1, id = 0, ..., tag)

release_all(.trj, resource, ..., tag)

release_selected_all(.trj, id = 0, ..., tag)

Arguments

.trj	the trajectory object.
resource	the name of the resource.
amount	the amount to seize/release, accepts either a numeric or a callable object (a function) which must return a numeric.
continue	a boolean (if post.seize OR reject is defined) or a pair of booleans (if post.seize AND reject are defined; if only one value is provided, it will be recycled) to indicate whether these subtrajectories should continue to the next activity in the main trajectory.
post.seize	an optional trajectory object which will be followed after a successful seize.
reject	an optional trajectory object which will be followed if the arrival is rejected (dropped).
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.
id	selection identifier for nested usage.

Details

Rejection happens when a resource is at full capacity and there is no room in the queue (either because there is a finite queue_size and it is full, or because queue_size=0 and thus it is disabled). In those cases, the reject parameter defines a fallback trajectory. Note, however, that, if the arrival is accepted (either in the queue or in the server) and then it is dropped afterwards due to preemption or resource shrinkage, then this trajectory will not be executed. Instead, see handle_unfinished for another, more general, method for handling all kinds of unfinished arrivals.

Value

Returns the trajectory object.

See Also

select, set_capacity, set_queue_size, set_capacity_selected, set_queue_size_selected

Examples

## simple seize, delay, then release
traj <- trajectory() %>%
  seize("doctor", 1) %>%
  timeout(3) %>%
  release("doctor", 1)

simmer() %>%
  add_resource("doctor", capacity=1) %>%
  add_generator("patient", traj, at(0, 1)) %>%
  run() %>%
  get_mon_resources()

## arrival rejection (no space left in the queue)
traj <- trajectory() %>%
  log_("arriving...") %>%
  seize("doctor", 1) %>%
  # the second patient won't reach this point
  log_("doctor seized") %>%
  timeout(5) %>%
  release("doctor", 1)

simmer() %>%
  add_resource("doctor", capacity=1, queue_size=0) %>%
  add_generator("patient", traj, at(0, 1)) %>%
  run() %>% invisible

## capturing rejection to retry
traj <- trajectory() %>%
  log_("arriving...") %>%
  seize(
    "doctor", 1, continue = FALSE,
    reject = trajectory() %>%
      log_("rejected!") %>%
      # go for a walk and try again
      timeout(2) %>%
      log_("retrying...") %>%
      rollback(amount = 4, times = Inf)) %>%
  # the second patient will reach this point after a couple of walks
  log_("doctor seized") %>%
  timeout(5) %>%
  release("doctor", 1) %>%
  log_("leaving")

simmer() %>%
  add_resource("doctor", capacity=1, queue_size=0) %>%
  add_generator("patient", traj, at(0, 1)) %>%
  run() %>% invisible

## combining post.seize and reject
traj <- trajectory() %>%
  log_("arriving...") %>%
  seize(
    "doctor", 1, continue = c(TRUE, TRUE),
    post.seize = trajectory("admitted patient") %>%
      log_("admitted") %>%
      timeout(5) %>%
      release("doctor", 1),
    reject = trajectory("rejected patient") %>%
      log_("rejected!") %>%
      seize("nurse", 1) %>%
      timeout(2) %>%
      release("nurse", 1)) %>%
  # both patients will reach this point, as continue = c(TRUE, TRUE)
  timeout(10) %>%
  log_("leaving...")

simmer() %>%
  add_resource("doctor", capacity=1, queue_size=0) %>%
  add_resource("nurse", capacity=10, queue_size=0) %>%
  add_generator("patient", traj, at(0, 1)) %>%
  run() %>% invisible

---

Select Resources

Description

Activity for selecting a resource for a subsequent seize/release or setting its parameters (capacity or queue size). Resources must be defined in the simulation environment (see add_resource).

Usage

select(.trj, resources, policy = c("shortest-queue",
  "shortest-queue-available", "round-robin", "round-robin-available",
  "first-available", "random", "random-available"), id = 0, ..., tag)

Arguments

.trj	the trajectory object.
resources	one or more resource names, or a callable object (a function) which must return one or more resource names.
policy	if resources is a character vector, this parameter determines the criteria for selecting a resource among the set of policies available (see details).
id	selection identifier for nested usage.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.

Details

The 'shortest-queue' policy selects the least busy resource; 'round-robin' selects resources in cyclical order; 'first-available' selects the first resource available, and 'random' selects a resource randomly.

All the 'available'-ending policies ('first-available', but also 'shortest-queue-available', 'round-robin-available' and 'random-available') check for resource availability (i.e., whether the capacity is non-zero), and exclude from the selection procedure those resources with capacity set to zero. This means that, for these policies, an error will be raised if all resources are unavailable.

Value

Returns the trajectory object.

See Also

seize_selected, release_selected, set_capacity_selected, set_queue_size_selected

Examples

## predefined policy
traj <- trajectory() %>%
  select(paste0("doctor", 1:3), "round-robin") %>%
  seize_selected(1) %>%
  timeout(5) %>%
  release_selected(1)

simmer() %>%
  add_resource("doctor1") %>%
  add_resource("doctor2") %>%
  add_resource("doctor3") %>%
  add_generator("patient", traj, at(0, 1, 2)) %>%
  run() %>%
  get_mon_resources()

## custom policy
env <- simmer()
res <- paste0("doctor", 1:3)

traj <- trajectory() %>%
  select(function() {
    occ <- get_server_count(env, res) + get_queue_count(env, res)
    res[which.min(occ)[1]]
  }) %>%
  seize_selected(1) %>%
  timeout(5) %>%
  release_selected(1)

for (i in res) env %>%
  add_resource(i)
env %>%
  add_generator("patient", traj, at(0, 1, 2)) %>%
  run() %>%
  get_mon_resources()

---

Inter-arrival Communication

Description

These activities enable asynchronous programming. send() broadcasts a signal or a list of signals. Arrivals can subscribe to signals and (optionally) assign a handler with trap(). Note that, while inside a batch, all the signals subscribed before entering the batch are ignored. Upon a signal reception, the arrival stops the current activity and executes the handler (if provided). Then, the execution returns to the activity following the point of the interruption. untrap() can be used to unsubscribe from signals. wait() blocks until a signal is received.

Usage

send(.trj, signals, delay = 0, ..., tag)

trap(.trj, signals, handler = NULL, interruptible = TRUE, ..., tag)

untrap(.trj, signals, ..., tag)

wait(.trj, ..., tag)

Arguments

.trj	the trajectory object.
signals	signal or list of signals, accepts either a string, a list of strings or a callable object (a function) which must return a string or a list of strings.
delay	optional timeout to trigger the signals, accepts either a numeric or a callable object (a function) which must return a numeric.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.
handler	optional trajectory object to handle a signal received.
interruptible	whether the handler can be interrupted by signals.

Value

Returns the trajectory object.

See Also

renege_if

Examples

## block, signal and continue with a handler
signal <- "you shall pass"

t_blocked <- trajectory() %>%
  trap(
    signal,
    trajectory() %>%
      log_("executing the handler")) %>%
  log_("waiting...") %>%
  wait() %>%
  log_("continuing!")

t_signaler <- trajectory() %>%
  log_(signal) %>%
  send(signal)

simmer() %>%
  add_generator("blocked", t_blocked, at(0)) %>%
  add_generator("signaler", t_signaler, at(5)) %>%
  run() %>% invisible

## handlers can be interrupted, unless interruptible=FALSE
t_worker <- trajectory() %>%
  trap(
  signal,
  handler = trajectory() %>%
    log_("ok, I'm packing...") %>%
    timeout(1)) %>%
  log_("performing a looong task...") %>%
  timeout(100) %>%
  log_("and I'm leaving!")

simmer() %>%
  add_generator("worker", t_worker, at(0)) %>%
  add_generator("signaler", t_signaler, at(5, 5.5)) %>%
  run() %>% invisible

---

Set Attributes

Description

Activity for modifying attributes. Attributes defined with set_attribute are per arrival, meaning that each arrival has its own set of attributes, not visible by any other one. On the other hand, attributes defined with set_global are shared by all the arrivals in the simulation.

Usage

set_attribute(.trj, keys, values, mod = c(NA, "+", "*"), init = 0, ...,
  tag)

set_global(.trj, keys, values, mod = c(NA, "+", "*"), init = 0, ..., tag)

Arguments

.trj	the trajectory object.
keys	the attribute name(s), or a callable object (a function) which must return attribute name(s).
values	numeric value(s) to set, or a callable object (a function) which must return numeric value(s).
mod	if set, values modify the attributes rather than substituting them.
init	initial value, applied if mod is set and the attribute was not previously initialised. Useful for counters or indexes.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.

Details

Attribute monitoring is disabled by default. To enable it, set mon=2 in the corresponding source (see, e.g., add_generator). Then, the evolution of the attributes during the simulation can be retrieved with get_mon_attributes. Global attributes are reported as unnamed key/value pairs.

Value

Returns the trajectory object.

See Also

get_attribute, get_global, timeout_from_attribute, timeout_from_global

Examples

env <- simmer()

traj <- trajectory() %>%

  # simple assignment
  set_attribute("my_key", 123) %>%
  set_global("global_key", 321) %>%

  # more than one assignment at once
  set_attribute(c("my_key", "other_key"), c(5, 64)) %>%

  # increment
  set_attribute("my_key", 1, mod="+") %>%

  # assignment using a function
  set_attribute("independent_key", function() runif(1)) %>%

  # assignment dependent on another attribute
  set_attribute("dependent_key", function()
    ifelse(get_attribute(env, "my_key") <= 0.5, 1, 0))

env %>%
  add_generator("dummy", traj, at(3), mon=2) %>%
  run() %>%
  get_mon_attributes()

---

Set Resource Parameters

Description

Activities for dynamically modifying a resource's server capacity or queue size, by name or a previously selected one. Resources must be defined in the simulation environment (see add_resource).

Usage

set_capacity(.trj, resource, value, mod = c(NA, "+", "*"), ..., tag)

set_capacity_selected(.trj, value, id = 0, mod = c(NA, "+", "*"), ..., tag)

set_queue_size(.trj, resource, value, mod = c(NA, "+", "*"), ..., tag)

set_queue_size_selected(.trj, value, id = 0, mod = c(NA, "+", "*"), ...,
  tag)

Arguments

.trj	the trajectory object.
resource	the name of the resource.
value	numeric value to set, or a callable object (a function) which must return a numeric value.
mod	if set, values modify the attributes rather than substituting them.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.
id	selection identifier for nested usage.

Value

Returns the trajectory object.

See Also

select, seize, release, seize_selected, release_selected, get_capacity, get_queue_size

Examples

## a resource with a queue size equal to the number of arrivals waiting
traj <- trajectory() %>%
  set_queue_size("res", 1, mod="+") %>%
  seize("res") %>%
  set_queue_size("res", -1, mod="+") %>%
  timeout(10) %>%
  release("res")

simmer() %>%
  add_resource("res", 1, 0) %>%
  add_generator("dummy", traj, at(0:2)) %>%
  run() %>%
  get_mon_resources()

---

Set Prioritization Values

Description

Activity for dynamically modifying an arrival's prioritization values. Default prioritization values are defined by the source (see add_generator, add_dataframe).

Usage

set_prioritization(.trj, values, mod = c(NA, "+", "*"), ..., tag)

Arguments

.trj	the trajectory object.
values	expects either a vector/list or a callable object (a function) returning a vector/list of three values c(priority, preemptible, restart). A negative value leaves the corresponding parameter unchanged. See add_generator for more information about these parameters.
mod	if set, values modify the attributes rather than substituting them.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.

Value

Returns the trajectory object.

See Also

get_prioritization

Examples

traj <- trajectory() %>%

  # static values
  set_prioritization(c(3, 7, TRUE)) %>%

  # increment
  set_prioritization(c(2, 1, 0), mod="+") %>%

  # dynamic, custom
  set_attribute("priority", 3) %>%
  set_prioritization(function() {
    prio <- get_prioritization(env)
    attr <- get_attribute(env, "priority")
    c(attr, prio[[2]]+1, FALSE)
  })

---

Set Source Parameters

Description

Activities for modifying a source's trajectory or source object by name. Sources must be defined in the simulation environment (see add_generator, add_dataframe).

Usage

set_trajectory(.trj, sources, trajectory, ..., tag)

set_source(.trj, sources, object, ..., tag)

Arguments

.trj	the trajectory object.
sources	the name(s) of the source(s) or a function returning the name(s).
trajectory	the trajectory that the generated arrivals will follow.
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.
object	a function modelling the interarrival times (if the source type is a generator; returning a negative value or a missing value stops the generator) or a data frame (if the source type is a data source).

Value

Returns the trajectory object.

See Also

activate, deactivate

Examples

traj1 <- trajectory() %>%
  timeout(1)

traj2 <- trajectory() %>%
  set_source("dummy", function() 1) %>%
  set_trajectory("dummy", traj1) %>%
  timeout(2)

simmer() %>%
  add_generator("dummy", traj2, function() 2) %>%
  run(6) %>%
  get_mon_arrivals()

---

Create a Simulator

Description

This method initialises a simulation environment.

Usage

simmer(name = "anonymous", verbose = FALSE, mon = monitor_mem(),
  log_level = 0)

Arguments

name	the name of the simulator.
verbose	enable showing activity information.
mon	monitor (in memory by default); see monitor for other options.
log_level	debugging level (see log_).

Value

Returns a simulation environment.

See Also

Available methods by category:

• Simulation control: stepn, run, now, peek, reset

• Resources: add_resource, get_resources, get_capacity, get_capacity_selected, get_queue_size, get_queue_size_selected, get_server_count, get_server_count_selected, get_queue_count, get_queue_count_selected, get_seized, get_seized_selected, get_activity_time, get_activity_time_selected, get_selected

• Sources: add_generator, add_dataframe, get_sources, get_n_generated, get_trajectory

• Arrivals: get_name, get_start_time, get_attribute, get_prioritization, get_batch_size

• Globals: add_global, get_global

• Data retrieval: get_mon_arrivals, get_mon_attributes, get_mon_resources

Examples

## a simple trajectory that prints a message
t0 <- trajectory("my trajectory") %>%
  log_("arrival generated")

## create an empty simulation environment
env <- simmer("SuperDuperSim")
env

## add a generator and attach it to the trajectory above
env %>% add_generator("dummy", t0, function() 1)

## run for some time
env %>% run(until=4.5)
env %>% now()           # current simulation time
env %>% peek()          # time for the next event
env %>% stepn()         # execute next event

---

Delay

Description

Activity for inserting delays and execute user-defined tasks.

Usage

timeout(.trj, task, ..., tag)

timeout_from_attribute(.trj, key, ..., tag)

timeout_from_global(.trj, key, ..., tag)

Arguments

.trj	the trajectory object.
task	the timeout duration supplied by either passing a numeric or a callable object (a function) which must return a numeric (negative values are automatically coerced to positive).
...	unused.
tag	activity tag name to perform named rollbacks (see rollback) or just to better identify your activities.
key	the attribute name, or a callable object (a function) which must return the attribute name.

Value

Returns the trajectory object.

See Also

set_attribute, set_global

set_attribute, set_global

Examples

env <- simmer()

traj <- trajectory() %>%

  # static delay
  timeout(3) %>%

  # dynamic, exponential delay
  timeout(function() rexp(1, 10)) %>%

  # dependent on an attribute
  set_attribute("delay", 2) %>%
  set_global("other", function() rexp(1, 2)) %>%
  timeout_from_attribute("delay") %>%
  timeout_from_global("other")

env %>%
  add_generator("dummy", traj, at(0)) %>%
  run() %>%
  get_mon_arrivals()

---

Create a Trajectory

Description

This method initialises a trajectory object, which comprises a chain of activities that can be attached to a generator. See below for a complete list of available activities by category.

Usage

trajectory(name = "anonymous", verbose = FALSE)

Arguments

name	the name of the trajectory.
verbose	enable showing additional information.

Value

Returns an environment that represents the trajectory.

See Also

Available activities by category:

• Debugging: log_, stop_if

• Delays: timeout, timeout_from_attribute, timeout_from_global

• Arrival properties: set_attribute, set_global, set_prioritization

• Interaction with resources: select, seize, release, release_all, seize_selected, release_selected, release_selected_all, set_capacity, set_queue_size, set_capacity_selected, set_queue_size_selected

• Interaction with generators: activate, deactivate, set_trajectory, set_source

• Branching: branch, clone, synchronize

• Loops: rollback

• Batching: batch, separate

• Asynchronous programming: send, trap, untrap, wait

• Reneging: leave, handle_unfinished, renege_in, renege_if, renege_abort

Manage trajectories:

• Extract or Replace Parts of a Trajectory: Extract.trajectory

• Join Trajectories: join

• Number of Activities in a Trajectory: length.trajectory, get_n_activities

Examples

## create an empty trajectory
x <- trajectory("my trajectory")
x

## add some activities by chaining them
x <- x %>%
  log_("here I am!") %>%
  timeout(5) %>%
  log_("leaving!")
x

## join trajectories
x <- join(x, x)

## extract and replace
x[c(3, 4)] <- x[2]
x

---

Wrap a Simulation Environment

Description

This function extracts the monitored data from a simulation environment making it accessible through the same methods. Only useful if you want to parallelize heavy replicas (see the example below), because the C++ simulation backend is destroyed when the threads exit.

Usage

wrap(.env)

Arguments

.env	the simulation environment.

Value

Returns a simulation wrapper.

See Also

Methods for dealing with a simulation wrapper: get_mon_arrivals, get_mon_attributes, get_mon_resources, get_n_generated, get_capacity, get_queue_size, get_server_count, get_queue_count.

Examples

## Not run: 
library(parallel)

mm1 <- trajectory() %>%
  seize("server", 1) %>%
  timeout(function() rexp(1, 2)) %>%
  release("server", 1)

envs <- mclapply(1:4, function(i) {
  simmer("M/M/1 example") %>%
    add_resource("server", 1) %>%
    add_generator("customer", mm1, function() rexp(1, 1)) %>%
    run(100) %>%
    wrap()
})

## End(Not run)

---