Adaptive pricing in multi-agent supply chain markets using economic regimes

Today's complex supply chain markets require adaptive product pricing. We propose a product pricing approach which assumes a double-bounded log-logistic distribution to be underlying offer prices, the parameters of which are estimated in real-time using Radial Basis Function Networks, based on available information. The relations between price distributions and available information are dynamically modeled, using economic regimes (characterizing market conditions) and error terms (accounting for customer feedback). Given the parametric approximations of price distributions, acceptance probabilities are estimated using a closed-form mathematical expression, which is used to determine the price yielding a desired quota. We implement our novel approach in the MinneTAC agent and test it against a price-following approach in the TAC SCM game. When competing against world's leading TAC SCM agents, performance significantly improves; bid efficiency increases and profits more than double.