MODEL FORMULATION AND SOLUTION FRAMEWORK FOR THE OPTIMAL BREEDING PROBLEM IN THE AXIE INFINITY GAME

Axie Infinity (AI) is a popular game that introduced the concept of “play-to-earn”, allowing players to profit by playing and breeding Axies (creatures). This work presents a mathematical formulation for the AI Breeding Problem (AIBP) where the breeder buys Axies from the AI Marketplace, breeds pairs by paying a fee and generating offspring, and sells all the Axies. A key challenge is selecting breeding pairs, since thousands of different offspring and selling prices can be generated due to breeding randomness. Moreover, the number of breeding pairs grows quadratic with the number of Axies available. Thus, a two-phase framework is proposed to solve the AIBP. First, a customized Adaptive Large Neighborhood Search quickly finds profitable breeding pairs, then a solver determines the optimal breeding strategy using the profitable pairs. Validation using data from the AI Marketplace demonstrates the efficiency of the proposed approach in achieving near-optimal profits within reduced computational time.