shapley_numba

Numba-based computations for cooperative games.

shapley_numba is a high-performance Python library for computing Shapley values and Harsanyi dividends in cooperative game theory. It leverages Numba’s JIT compilation to achieve fast computations even for games with many players.

Features

Fast exact Shapley calculation - Optimized implementations using Numba
Shapley Monte Carlo approximation - For large-scale games
Harsanyi Dividends (synergies) computation - Analyze player interactions
Set iteration tools - Efficient subset enumeration utilities

Installation

Using pip:

pip install shapley-numba

Using uv:

uv add shapley-numba

Quick Example

from shapley_numba.shapley import shapley
from shapley_numba.examples import GloveGame

glove_game = GloveGame(num_left_gloves=1)
shapley(glove_game, num_players=3)