The Innovation Game
_{The Network for Algorithmic Breakthroughs}

The Innovation Game (TIG) creates a new economic framework for algorithmic development - one that aligns incentives, rewards contribution, and keeps innovation open.

At its core, TIG uses a novel proof-of-work variant built around computational challenges grounded in scientifically important problems. Innovators submit algorithms that solve these challenges, and benchmarkers are incentivized to adopt the most efficient ones for proof-of-work, creating a manipulation-resistant signal for rewarding the top-performing algorithms.

In this way, TIG democratizes algorithmic innovation, turning contribution into a sustainable economic opportunity, coordinating global intelligence to compete with centralized incumbents.

Challenges

TIG currently has 8 active computational challenges:

ID	Challenge	Description	CPU/GPU
c001	satisfiability	Boolean Satisfiability (SAT)	CPU
c002	vehicle_routing	Capacitated Vehicle Routing with Time Windows	CPU
c003	knapsack	Quadratic Knapsack Problem	CPU
c004	vector_search	Vector Range Search	GPU
c005	hypergraph	Hypergraph Partitioning	GPU
c006	neuralnet_optimizer	Neural Network Optimizer	GPU
c007	job_scheduling	Flexible Job Shop Scheduling	CPU
c008	energy_arbitrage	Energy Market Arbitrage	CPU

Glossary

Term	Definition
Innovator	Participant who submits algorithms (code or advances) to solve challenges
Benchmarker	Participant who runs algorithm benchmarks and submits proofs
Challenge	A computational problem adapted for optimisable proof-of-work
Code	An algorithm source code submission by an Innovator
Advance	An algorithm improvement submission (documentation/paper) by an Innovator
Fuel	Computational cost metric — algorithms must solve within a fuel budget
OPoW	Optimisable Proof of Work — TIG's core consensus mechanism
Nonce	Input seed for a single benchmark run
Runtime Signature	Hash produced during algorithm execution, used for verification
Binary	A compiled shared object (.so) built from an algorithm submission

Important Links

• Getting Started for Innovators
• Getting Started for Benchmarkers
• TIG Documentation
• TIG Whitepaper
• TIG Licensing Explainer
• Code vs Advances
• Voting Guidelines for Token Holders

Repo Contents

Crate	Description
tig-algorithms	Hosts algorithm submissions (code and advances) made by Innovators
tig-benchmarker	Python scripts for running TIG's benchmarker in master/slave configuration
tig-binary	Wraps an algorithm submission for compilation into a shared object
tig-challenges	Implementations of TIG's 8 computational challenges
tig-protocol	Core protocol logic (block processing, submissions, verification)
tig-runtime	Executes a compiled algorithm for a single nonce, generating runtime signature and fuel consumed
tig-structs	Shared struct definitions used throughout TIG
tig-token	Solidity ERC20 token contract deployed on Ethereum L2 Base chain
tig-utils	Utility functions (hashing, Merkle trees, serialization, etc.)
tig-verifier	Verifies a single solution or Merkle proof

Docker Images

TIG Docker images are hosted on GitHub Packages, supporting linux/arm64 and linux/amd64 platforms.

Note: Check tig-benchmarker/.env for the current VERSION (currently 0.0.5).

Dev Images (for Innovators)

Development environment for writing and compiling algorithms:

Challenge	Image
satisfiability	satisfiability/dev
vehicle_routing	vehicle_routing/dev
knapsack	knapsack/dev
vector_search	vector_search/dev
hypergraph	hypergraph/dev
neuralnet_optimizer	neuralnet_optimizer/dev
job_scheduling	job_scheduling/dev
energy_arbitrage	energy_arbitrage/dev

Slave Images (for Benchmarkers)

Runtime images spun up as part of slave.yml (see benchmarker README):

Component	Image
Slave orchestrator	benchmarker/slave
satisfiability	satisfiability/runtime
vehicle_routing	vehicle_routing/runtime
knapsack	knapsack/runtime
vector_search	vector_search/runtime
hypergraph	hypergraph/runtime
neuralnet_optimizer	neuralnet_optimizer/runtime
job_scheduling	job_scheduling/runtime
energy_arbitrage	energy_arbitrage/runtime

Master Images (for Benchmarkers)

Spun up as part of master.yml (see benchmarker README):

Component	Image
Master orchestrator	benchmarker/master
Dashboard UI	benchmarker/ui
PostgreSQL	benchmarker/postgres
Nginx reverse proxy	benchmarker/nginx

Useful Scripts

The runtime and dev images include these scripts on PATH:

list_algorithms                              # List available algorithms for the challenge
download_algorithm <algorithm_name_or_id>    # Download an algorithm's source
test_algorithm <algorithm_name> <difficulty>  # Test an algorithm locally

The container automatically sets the CHALLENGE environment variable (e.g. knapsack/runtime sets CHALLENGE=knapsack). Use --testnet to target testnet.

License

See README for individual folders.