Distribution and agentification of heuristic methods applied to complex systems in AI: a review of the literature

Abstract

In this paper, we presented a review of the literature on the application of distributed and agentified heuristic methods to complex systems in AI. We begin by introducing the key concepts of distribution and agentification, and the benefits and challenges associated with their use in complex systems. Then, we review some heuristic approaches that have been developed to solve complex problems in areas such as optimization, planning, pattern recognition, learning, etc. In particular, we focus on heuristic methods that can be distributed and/or agentic, highlighting their main features and the results obtained in a concrete application. Finally, we discuss the future prospects and challenges of distributing and agenticating heuristic methods in AI.