Mathematics for price negotiation

Mathematics: game theory, Markov chains with hidden states, multi-agent modeling

Electis is a distribution company in electronic and energy equipment. It is the link between manufacturers (ex: Legrand, Hager, Schneider, etc ...) and professional customers (installers electricians, industrial, administrations, etc ...). A distributor deals with logistics, storage, after-sales service, and other services. For the good functioning of the company, the company must make profits between the purchase of the equipment at the manufacturers and the sale of the equipment to the customers. For Electis, the clientele consists exclusively of professionals. In this context, we encounter several difficulties (i) Competitor prices are unknown, there is no "reference price" or market price for these products(except for a public price) (ii) the customer typology is heterogeneous from the small craftsman to very large groups.

As part of the SEME, it is proposed to model a distributor's business and strategic environment, as well as interactions with competing manufacturers, customers, and distributors based on a reduced set of dummy data generated by using the distributions of true data. It is a question of studying if the modeling makes it possible to find the interactions of negotiations in the lines of sale. We can think of different types of mathematics to deal with the problem: (i) game theory, (ii) hidden state Markov models, or (iii) multi-agent modeling.

The goal of SEME on this subject is to provide some ideas for modeling and algorithmic thinking to deal with the subject. Electis is particularly interested in (i) establishing a state of the art on the basis of preliminary work on mathematical modeling of the negotiation and the related algorithms and (ii) exploring one of the possible paths of negotiation modeling through data. available.

More details here

Note: The Electis project will eventually be split into two projects depending on the type of approach chosen.