Agent-Based Evolutionary Game-Theoretic Framework for Optimizing Investment Plans of Competitive Port Facilities

Abstract

Abstract The current paper deals with a special case of the Network Design Problem (NDP) that arises when multiple authorities are controlling certain parts of a network and in particular when these are engaged into a deliberative environment, forming a competitive market for transportation services. This alteration departs significantly from the classical paradigm of the NDP extending the classical single leader-multiple followers Stackelberg game-theoretic formulation to its complete form of multiple leaders-multiple followers Competitive NDP (CNDP). In order to address the particular problem setup, a multilevel vector-optimization programming problem is formulated, which is tackled by a proposed novel Artificial Intelligence (AI)-based decentralized evolutionary optimization algorithm. The application of this framework is done over a freight network that is composed of multiple transportation means, and in particular over a complete transportation chain formed by shippers, maritime and land carriers and port authorities, where the latter are competing for profits. The market of maritime facilities is modeled as an n-person non-cooperative game among port authorities who control the attractiveness of their terminal facilities. By taking the above interdependencies into consideration, the estimation of the equilibrium point of the above formulation is made by creating a game-theoretic platform where Intelligent Agents (corresponding to port authorities) are evolving their strategies towards equilibrium points. Results from the application of the proposed framework into a realistic part of the East Mediterranean freight network are provided and discussed.