A multilateral multi-issue automated negotiation model based on NSGA-III

Abstract:
Compared with bilateral negotiations, agent-based negotiation model design is more complicated in multilateral negotiations. However, most of the existing multilateral negotiations are essentially one-to-one concurrent bilateral negotiations. This method usually performs with low transaction efficiency and high cost and cannot meet the complicated scenario requirements, in which multiple agents seek a consensus agreement. To solve this problem, we propose a multilateral multi-issue automated negotiation (MMN) model based on the NSGA-III algorithm. First, we transform the MMN problem into a multi-objective optimization model, so as to characterize the negotiation process on the Pareto front in a multi-dimensional space. Second, we propose a negotiation mechanism suitable for MMN, which can not only solve one-to-many simultaneous negotiation problems, such as multilateral procurement and distribution, but also solve multi-agent simultaneous negotiation problems, such as resource (interest) allocation. The case study results show that the proposed model and algorithm can effectively solve the MMN problems and are helpful to achieve a multilateral win-win situation.