Apartment Rental Benchmarking (ARB) MCMC simulations of apartment rent finance and outcomes Linux / R programming language exercise (C) Jukka Paulin 2012
This memo drafts both a real-world simulator for apartment renting and a computer game derived from the idea(s).
The ARB is a simulation of rent-letting, in the environment of a Market (of houses). The activity takes place in iterations, or rounds. With each individual round only a handful of different actions can happen. The simulation can accomodate to two main modes: without a "letting agent" (the landlord acts solo), or using an agent as a facilitator between the landlord and the tenant.
A discrete MCMC simulation model is used. This means essentially that (a) the simulation advances in steps (='it is discrete') (b) Monte Carlo
The simulation is centered around a single house for rental (the Apartment). The house can be occupied or free; if it is occupied, the owner gets a sum of money each month (assuming the tenant is honest).
The simulation branches as time advances, and there are interesting end-points where the agents of this simulation have experienced different "train" of events. Randomness (luck) has a saying in the outcome of the simulation, as does other factors.
The simplified version of this program just loops through N iterations and spills out the results. This version contains (7) of the basic nine events. The neglected events are those two that have to do with using couriers as middlemen in the renting business.
The real beef of the simulation is scenario analysis. Scenarios can be made by tweaking static variables or touching the limits of random variables.
Function definitions
The ARB simulates how rental activity (owning a house and renting it directly or indirectly via couriers) takes different kinds of paths according to events.
The events can be one of the (9) possible:
- seeking a new candidate to occupy the rental house
- forming or purging an rental agreement
- the candidate purges a rental agreement
- the rental agreement is expired and not continued
- the rental agreement expires and is continued
- forming or purging a courier agreement
- selling the property Advanced events include (2) more:
- candidate neglects the agreement by not paying rent
- candidate neglects the agreement and causes damage to apartment
- data structures are quite simple
- things to keep record of are
- each house status:
- tenant currently, or None
- if None, then a code that signifies one of Nine Eventcodes
- each house status:
Markets
- demand
- reasons for demand
Population density and other interesting statistics
- square kilometers used as areal unit
We keep notice of the vacancy of the apartment. This is stored in the variable Y. Y can have only two states: 0 = empty, 1 = vacant
- qualitative is about the "human" side of an issue; often contradictory behaviors, beliefs, opinions, emotions and relationships of individuals
- qualitative research is also effective in identifying intangible factors, such as social norms, socioeconomic status, ...