Receding Horizon Planning for Economic Coordination

Python-implementation of receding horizon planning technique for macroeconomic coordination of production. Code is readily used together with real economic data in the form of supply and use tables.

Background

In the era of climate change, viable societies will have to rapidly adapt to extreme weather events as well as coordinate economic activities in an efficient manner towards long-term goals. This will require a careful design and experimentation of coordination protocols and optimization procedures for large-scale economic systems.

The code in this repository implements a type of optimization procedure which we call receding horizon planning (RHP). It builds on ideas from optimization theory, economic analysis and control theory (cf. the works of Kantorovich, von Neumann, Bellman and Bertsekas). An introductory technical description is provided here.

A first example with Swedish national accounts data is given here.

Python implementation

The goal is to coordinate $m$ units of production, so that their production levels utilize the minimal amount of resources required to meet material constraints over a future time horizon. This coordination process is revised at every time step.

Variables

Note: full implies that exports are included, they are otherwise omitted from the target output in question.

TeX paper	Code	Description	Type & length
$T$	time_steps	simulated time steps	non-negative integer & 1
$N$	planning_horizon	planning horizon	non-negative integer & 1
$c$	primary_resource_list	it might be worked hours or co2 for example	list, each item in the list numpy.matrix & n*1
$J$	supply_use_list	supply minus use list in price unit	list, each item in the list numpy.matrix & n*m
$\bar{T}$	use_imported_list	use imported list in product price unit	list, each item in the list numpy.matrix & n*m
$D$	depreciation_matrix_list	depreciation matrix list	list, each item in the list numpy.matrix & n*n
$r$	full_domestic_target_output_list	domestic production target in product price units	list, each item in the list numpy.matrix & n*1
	export_constraint_boolean	a boolean being true if the export constraint is used	boolean
$r_{exp}$	augmented_export_vector_list	export vector in product price unit	list, each item in the list numpy.matrix & m*1
$p_{exp}$	export_prices_list	export prices list	list, each item in the list numpy.matrix & m*1
$p_{imp}$	import_prices_list	import prices list	list, each item in the list numpy.matrix & m*1
	upper_bound_on_activity	sets an upper bound on activity	None or int
	max_iterations	sets an upper bound on iteration number	int
	tolerance	sets a tolerance for what constitutes a solution	int