Skip to content

An NPM package to assign groups / persons to events based on their preferences

License

Notifications You must be signed in to change notification settings

Altesmi/eventassigner-js

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

30 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

eventassigner-js

An NPM package to assign groups / persons to events based on their preference.

Uses modified algorithm from Li et al. (2014)

The algorithm is modified to handle groups of varying size.

An observable notebook where the new algorithm is tested against real data is located here

How to use version 0.1.*

Install the package

npm install --save eventassigner-js

Require the package

const ea = require('eventassigner-js')

Define event assignment problem

function ea.eventAssignment(input) takes an object which has to following fields

groups: array of objects whose each entry must have the following fields

  1. id: unique identifier
  2. size: size of the group. Must be a number
  3. pref: Array whose elements maps to event identifiers. Must be at least an empty array.

events: array of objects whose each entry must have the following fields

  1. id: unique identfier. -1 can not be an id
  2. min: minimum number of attendees the event needs for it to happen. Must be a number.
  3. max: maximum number of attendees the event can hold. Must be a number.
  4. groups: groups that are attending to this event. Must be at least an empty array.

list: array of objects whose each entry must have the following fields.

  1. id: id of a group
  2. size: size of the group
  3. event: id of the event this group wants to participate
  4. gain: numeric value which describes how much the algorithm gains if this group is assigned to this event.

This list is traversed in the optimization starting from the last index and moving towards the first index with Array.prototype.pop() function. At least currently the algorithm does not sort the list. Any list element with gain: 0 is removed before any optimization happens. If you want to maximize the gain in the list it is wise to sort the list array so that the entries with highest gain are at the bottom of the list (see example of the sorting below in the 'working example' section)

updateL: a function which takes an object as a input and returns a new list. This function is called every time in the loop when an entry in 'list' is not assigned to the event specified by that entry. The function can be used to update the gains in 'list' during the optimization. The input to the function includes following fields

  1. groups: same as the input

  2. events, same as the input

  3. assignment: current assignment which has the following fields

    • id: group id
    • assignment: the events id where this group is assigned. -1 means no assignment
  4. unassignedGroups: groups that are not assigned. Has the same structure as 'groups' array

  5. deficit: total number of attendees needed for those events to happen that have at least one attendee but less than the minimum amount of attendees needed. Defined as

  6. groupId: the id of the group that was not assigned

    equation

    where

    • P: events that have at least one attendee but less than the minimum amount of that event
    • equation: minimum amount of attendees needed for event a to happen
    • |E_a|: number of attendees in event a
  7. list: current 'list' array

updateL()-function must return a list object. Minimum working function is

const updateL = input => input.list

Call the eventAssignment function

const assignment = ea.eventAssignment(input)

Evaluate output

Upon successful run, eventAssignment() function returns the assignment array whose every entry has following fields

  1. id: group id
  2. assignment: the events id where this group is assigned. -1 means no assignment

The function returns 0 if the input is not correct.

If the input is correct but for some reason the assignment is not correct, the return value is an object whose fields are

  1. result: 0 i.e. the assignment is not correct.
  2. events: events that have either less than minimum number of players assigned to them and are not happening or events that have more than maximum number of players assigned to them.
  3. flag: string which describes what went wrong.

The assignment can be wrong if

  1. There is less than minimum number of players assigned to any event.
  2. There is more than max number of players assigned to any event.

Working example

const ea = require('eventassigner-js');

const groups = [{
  id: 1,
  size: 2,
  pref: [1, 2, 3],
},
{
  id: 2,
  size: 1,
  pref: [2, 3],
},
{
  id: 3,
  size: 2,
  pref: [3],
},
]

const events = [{
  id: 1,
  min: 1,
  max: 3,
  groups: [],
},
{
  id: 2,
  min: 1,
  max: 1,
  groups: [],
},
{
  id: 3,
  min: 1,
  max: 3,
  groups: [],
},
]

let L = [
  {
    id: 1,
    size: 2,
    event: 1,
    gain: 1,
  },
  {
    id: 1,
    size: 2,
    event: 2,
    gain: 0.5,
  },
  {
   id: 1,
    size: 2,
    event: 3,
    gain: 0.33,
  },
  {
    id: 2,
    size: 1,
    event: 2,
    gain: 1,
  },
  {
    id: 2,
    size: 1,
    event: 3,
    gain: 0.5,
  },
  {
   id: 3,
    size: 2,
    event: 3,
    gain: 1,
  },
]

L = L.sort((a, b) => {
  if (a.gain >= b.gain) {
    return 1
  } else {
    return -1
  }
})

const updateL = input => input.list

const input = {groups,
events,
list: L,
updateL}

const assignment = ea.eventAssignment(input)

console.log(assignment)

Saving and running above example

node workingexample.js
[ { id: 1, assignment: 1 },
  { id: 2, assignment: 2 },
  { id: 3, assignment: 3 } ]

About

An NPM package to assign groups / persons to events based on their preferences

Resources

License

Stars

Watchers

Forks

Packages

No packages published