A data-driven design of the optimal investment portfolio for the industry in a two-level game using the Markowitz model by meta-heuristic algorithms: Economic analysis of condition monitoring system

Abstract

This paper studies the investment portfolios of two players in the banking system in a two-level game and then, determines optimal portfolios of investors using the Markowitz model. This two-level game includes Bank C as the leader and customers of this bank as the game followers. The investment portfolios of the leader include investment in competitor banks (A and B), foreign exchange market, real estate market, and stocks. The data related to the mentioned assets covered the years 2010{2020, in which the optimal investment portfolios of the players were first determined using GAMS and genetic meta-heuristic algorithm. Next, the problem was solved again using the meta-heuristic algorithms of Particle Swarm Optimization (PSO) and Invasive Weed Optimization (IWO). Eventually, the optimal algorithm was chosen using TOPSIS multi-criteria decision-making. The results of the 3 algorithms indicated that the optimal portfolio for the leader player consisted of investment in properties, securities, and competitor banks.