Existence of optimal strategies in a normalized interval matrix game and applications

Abstract

Vagueness in the real-world problem performs a significant role in determining the existence of a feasible solution to the problem. This vagueness in the model may be handled by considering the parameters of the problem as a closed interval. In this competitive world, game models with uncertain payoffs can successfully handle conflicting real-world problems. Therefore, a two-player zero-sum game model in which payoffs vary in a range is considered. Then, the existence of saddle point (pure strategy) and dominance rules are discussed for the game model, and mixed strategy is defined in the case of non-existing saddle point. Thereafter, a methodology has been developed to discuss the existence of weak and control strategies in the interval game model. Furthermore, we obtain the optimal value of the game model as a deterministic number. Finally, the developed methodology is successfully implemented to solve real-world investment problems in the stock market and obtain the optimal strategy to achieve the maximum return.

Long Short-Term Memory-Deep Belief Network-Based Gene Expression Data Analysis for Prostate Cancer Detection and Classification

“…The paper concludes by delving into a discourse on the future of machine learning, contemplating potential advancements in the field that might lead to the development of innovative systems. The objective of the presented research is to determine the most effective strategy utilizing state-of-the-art computer vision, specifically the Chaotic Oppositional Based Whale Optimization Algorithm (CO-WOA), in conjunction with data mining techniques [41,42,43,44].…”

A social image recommendation system based on deep reinforcement learning

“…Prior studies have explored handcrafted features [1,2], as well as deep learning models [3][4][5] for image representation and recommendation. Feature selection and the use of deep learning models have been prominent in many studies [6][7][8]. In recommender systems, various similarity measures [9] and classification methods [10] have been employed.…”