| The growing penetration of Distributed Energy Resources (DER) into distribution networks requires decentralized market mechanisms that remain physically feasible while using accurate day-ahead price forecasts to support local scheduling decisions. Inspired by mycorrhizal fungal networks, this dissertation introduces a bio-inspired, forecast-informed framework that integrates peer-to-peer energy trading, network-constrained coordination, and deep learning–based price prediction. The framework combines MYCO-MARKET, a decentralized sealed-bid double-auction market layer; SLIMER, a coordination layer that incorporates line, transformer, voltage, and loss price signals through a subgradient primaldual scheme with adaptive conductance weighting; and BOL-LPP, a Bayesian-optimized LSTM model for day-ahead price forecasting. The framework is evaluated through standalone forecasting, standalone market clearing, integrated network-constrained coordination, and full forecast-informed operation on radial and meshed distribution networks, including IEEE 33-, 37-, 69-, and 123-bus benchmark feeders. Results show that MYCO+SLIMER reduces community cost, eliminates trade-induced network violations in the principal tested scenarios, and approaches the centralized OPF benchmark under congested conditions. In parallel, BOL-LPP outperforms statistical, machine-learning, and deep-learning forecasting baselines across multiple Electric Reliability Council of Texas (ERCOT) pricing zones. The convergence and feasibility behavior of the coordination mechanism are characterized empirically across topologies and congestion levels, advancing decentralized, network-aware peer-to-peer energy trading with high-accuracy day-ahead forecasting. Keywords: P2P energy trading, network-constrained coordination, electricity price forecasting, MYCO-MARKET, SLIMER, BOL-LPP |