Multi-agent System Optimal Expansion Planning for Management of Renewable Distributed Generations

Abstract

Nigeria's centralized electrical infrastructure has regularly failed to satisfy the country's rising energy consumption, therefore impeding technical advancement and economic growth. Renewable Distributed Generations (RDGs) including solar PV, wind, and small hydro provide a sustainable route toward distributed energy access and enhanced grid dependability when combined. On operational stability, coordination, and growth planning, however, the variability and spatial dispersion of RDGs present difficult problems. Leveraging intelligent, autonomous agents representing smart microgrids, generating units, storage systems, and load centers, this work proposes a Multi-Agent System-based Optimal Expansion Planning (MAS-OEP) framework. These agents link long-term investment planning with real-time operational methods under uncertainty utilizing a two-stage stochastic optimization structure to generate distributed decisions. While scenario trees organize future situations, probabilistic models, Monte Carlo simulations, fuzzy sets, and Markov chains are used to reflect uncertainty in wind output, load demand, and market prices. The hybrid optimization system combines accurate methods such as Mixed- Integer Linear Programming (MILP) and Genetic Algorithms (GA) at the agent level with heuristic techniques at the upper level. Under general direction from a Central Coordination Agent (CCA), coordination is accomplished through peer-to--peer energy trading using the Contract Net Protocol (CNP) and Agent Communication Language (ACL). Furthermore, improved by the Genetic Vertical Sequencing Protocol (GVS) are system scalability, durability, and adaptability to evolving grid environments. While trade-off studies expose ideal balances between energy storage and backup generation, simulation findings demonstrate the efficacy of the concept in minimizing costs, increasing renewable usage, and reducing emissions. Thus, the suggested MAS-OEP architecture offers a scalable, intelligent, and strong method to control distributed energy expansion under uncertainty. It greatly raises operational dependability, environmental sustainability, and economic efficiency of Nigeria's power system. Future studies should investigate how to improve agent decision-making by means of real-time analytics, reinforcement learning, and adaptive regulatory models. Recommendations for full-scale deployment are strong policy support, secure communication infrastructure, and stakeholder capacity building. For Nigeria and other developing nations, this work provides a strong basis for improving smart, robust, and adaptive energy systems, simulation results demonstrate improved renewable utilization, reduced emissions, and enhanced grid stability, validating the robustness of the research framework.