MatchingMarket: Matching Market Class

In dymium-org/dymiumCore: A Toolkit for Building a Dynamic Microsimulation Model for Integrated Urban Modelling

Description

This is a the abstract base class for matching market objects like MatchingMarketStochastic and MatchingMarketOptimal. This class is usually used for solving problems such as allocation, matching and mating of agents.

Format

R6::R6Class object.

Details

sub-matching problems only contains agentsets' data in a named list format (agentset_A, agentset_B).

Construction

matching_problem <- MatchingMarket$new(agentset_A, agentset_B, slots_B, id_col_A, id_col_B, grouping_vars)

• agentset_A :: data.table::data.table()
  A data.table contains agentset A data and relavant attributes. This is usually the choosers (in one-sided matching).

• agentset_B :: data.table::data.table()
  A data.table contains agentset B data and relavant attributes. This is usually be the agents in the candidate pool.

• id_col_A :: character(1)
  A character that indicates the id column of agentset_A.

• id_col_B :: character(1)
  A character that indicates the id column of agentset_B.

• slots_B :: integer()
  (NULL)
  A integer vector of length 1 or length equals to the number of rows in agentset_B. This is needed in many-to-one matching. Such as workers to jobs.

• grouping_vars :: character()
  (NULL)
  When this is provided matching can be done just between agents of the same group.

Fields

• matching_problem
  

Methods

• split_market()
  (logical(1)) -> ()
  Create a list of list of data.tables. list({group_name} = list({sub_market_number} = list(agentset_A = data.table(), agentset_B = data.table()))).

• split_by_group(matching_problem = self$matching_problem, grouping_vars = self$matching_problem$grouping_vars)
  (agentset_A = data.table::data.table(), agentset_B = data.table::data.table()|logical(), character(), logical(1)) -> [data.table::data.table()]
  Create n-groups sub-matching problems.

• split_by_n(matching_problem = self$matching_problem, n = NULL)
  (agentset_A = data.table::data.table(), agentset_B = data.table::data.table()| integer()) -> list([data.table::data.table()]) Create n sub-matching problems. If n is not given, then n will be determine based on self$max_market_size.

• matching_score_A(matching_problem = self$matching_problem, idx_A, idx_B)
  (list(agentset_A = data.table::data.table(), agentset_B = data.table::data.table())|integer()|integer() ) -> integer() of length idx_B or matrix() of length idx_A * length idx_B
  This is an abstract function meaning it needs to be implemented. Please see the detail section for suggestion how this method should be implemented.

• matching_score_B(matching_problem = self$matching_problem, idx_B, idx_A)
  (list(agentset_A = data.table::data.table(), agentset_B = [data.table::data.table()]) |integer()|integer() ) -> integer() of length idx_B or matrix() of length idx_A * length idx_B
  This is an abstract function meaning it needs to be implemented. Please see the detail section for suggestion how this method should be implemented.

dymium-org/dymiumCore documentation built on July 18, 2021, 5:10 p.m.