sample_group_specification_annotated.groupeng

# Example input file for GroupEng
 
# Lines that start with a # are comments and GroupEng ignores them

# The file that containes the student list.  This should be a csv
# file as exported from excel
# Either store the input deck in the same directory as the class
# file,  or give the full path to the class file here
classlist : sample_class_1.csv

# Student Identifier (name, id number, ...) should be unique
student_identifier : ID

# group_size: number of students per group immediately followed by + or -
# If the students don't divide evenly into group_size person groups,
# group size can be + (have an extra person in some groups), or 
# - (have one less person in some groups)
group_size : 4+

# Alternatively, you can specify a number_of_groups to tell groupeng to make
# exactly that many groups and pick the group size automatically. Do not use
# both group_size and number_of_groups. If you do, your group_size will be ignored
# number_of_groups : 14

# Don't isolate women
- cluster : Gender
  values : M

# Don't isolate minorities
- cluster : Ethnicity
# a , b for a list where each value is clustered individually
# c = d for union of values, treat them both as one value,
# you can chain equals (c = d = f = ...) for larger unions
  values : B = H

# Put students on the same project choice together
- aggregate : Project choice

# Multidisciplinary teams
- distribute : Major
  values : Mech E, CS, Civ E, EE

# Ensure Teams have all needed skills
- distribute : Skill1
  value : y
- distribute : Skill2
  value : y
- distribute : Skill3
  value : y

# You will usually want to put the balance rule last.  It turn out the
# program is pretty good at meeting balance rules even with low
# priority, but locking in groups by gpa balance at high priority
# makes it hard to meet other rules
- balance : GPA

# you can specifiy a tol (tolerance) argument (below, commented out),
# but generally that is not necessary, the default value of .5
# generally works well, but tweaking this value may allow you to coax
# better balanced groups. It controls how many standard deviations the
# mean GPA of a group is allowed to be from the mean GPA of the class
# before the balance rule is considered "broken". Putting a tight
# (small number)tolerance balance early will remove most of GroupEng's
# freedom to operate and lead to bad groups. However, putting a tight
# tolerance late is generally safe, and will cause groupeng to try hard
# to make well balanced groups, though it may lead to many reports of
# the balance rule failing (because, while tightly balanced, they may not
# quite meet the tolerance specified. You can also apply multiple balance
# rules on the same attribute, one loose tolerance early to establish a
# required baseline for how balanced groups absolutely have to be and then
# a tighter one at the end to get groupeng to do any further optimization
# it can given the other rules
  tol : .2