Brussels, Belgium

Abstract:

This is a simple implementation example of the developed Spatial Microsimulation Urban Metabolism Model (SMUM).

The aim of this model is to identify and quantify the impact of transition pathways to a circular economy.

Two main algorithms implemented in the model, giving it it’s name, are:

  1. A Spatial Microsimulation, used for the construction of a synthetic population; and
  2. An Urban Metabolism approach, used to benchmark consumption level at a city-level or neighbourhood level (making it spatial).

(Step 1) Constructing a synthetic population

On this section the example shows how to construct representative samples given the distributions of aggregated variables. The algorithm constructs the samples either by constructing a new sample for each simulation year (resample method) or by reweighting an initial sample for each simulation year (reweighting method).

In order to construct the samples the model needs input on: (a) the distribution functions of aggregate variables and (b) the changes over time. This means that the model required the projected aggregated values. The model provides some function to facilitate and visualize the projection of aggregated values.

For a spatial microsimulation, the model requires also values for each simulation area. In this case a combination of the resample and reweighting methodologies is implemented. The algorithm will construct a new sample for each simulation year at an aggregated level (e.g. city-level) and reweight this sample for each area (e.g. statistical census areas).

For the computation of resource consumption the model requires a consumption model. This model defines how the resource values are computed. This can be anything from a simple linear model to the implementation of external libraries.

Aggregate level benchmarks

(Step 1.a) Projecting demographic variables

Micro level consumption models

(Step 1.c) Micro-level Electricity demand model

(Step 1.d) Micro-level Water demand model

(Step 1.e) Micro-level Non-Residential model

(Step 2) Sampling and reweighting

This section takes care of the actual sampling procedure and subsequent reweighting of the proxy data. The sample will be constructed with help of an MCMC algorithm and reweighted with help of the GREGWT algorithm.

This section also presents the internal validation of the sampling procedure.

Dynamic samplic models

(Step 2.a) Dynamic Sampling Model and GREGWT

(Step 2.b) Non-Residential Model

Model internal validation

(Step 2.c) Model Internal Validation

(Step 3) Constructing scenarions:

The construction of scenarios can be define at different steps of the model. In a sense, the definition of scenarios start by the projection of aggregated values (see section 1). This section defines scenarios by defining technology adoption rates and changes in technology efficiency.

(Step 3.a.1) Define Transition Scenarions

(Step 3.b.2) Visualize transition scenarios

(Step 3.a.2) Define Transition Scenarions, Non-Residential

(Step 3.b.2) Visualize transition scenarios, Non-Residential