Table 1 Comparison of selected existing studies.
Refs. | Technique/models | Contribution | Research gap | Proposed work (GrCRA-PCRTAM-Net) |
|---|---|---|---|---|
FDA-CNN for DC/AC MGs with EVs | Predictive control of EV charging and power sharing between DC/AC MGs | Limited to EV integration; lacks multi-objective optimization or hybrid storage management | Adds battery + SC management and multi-objective scheduling under RES variability | |
HOA for DG siting + network reconfiguration | Improves voltage profile and reduces losses | No forecasting integration; single optimization stage | Combines GrCRA optimization with forecast-guided scheduling | |
Hierarchical control (Sliding mode + Lyapunov) with NN + FLC | Improves PV + Battery + SC coordination in DC MG | Control-oriented; no predictive intelligence; limited to small MG | Integrates PCRTAM-Net prediction with optimization for larger distribution networks | |
PSO + PI control for HES (PV, FC, Battery, SC) | Reduces hydrogen use, balances HES operation | Focused on PI tuning; not scalable; limited forecasting | Embeds deep learning-based forecasting with scalable bio-inspired optimization | |
NSGA for multi-objective energy management | Balances cost, emissions, and reliability | Does not integrate hybrid storage or prediction | Provides SoC-aware hybrid storage control + predictive forecasting | |
SWO + MHFAN for PV-based EV charging | Predicts solar + demand, optimizes EV charging | Limited to EV charging applications; no general MG framework | Extends to general power distribution networks with integrated renewables | |
DFIG + PWM + Bird Swarm optimized PI for hybrid RES | Improves PQ and voltage in RES | Focused on converter-level control; not system-wide optimization | Addresses system-level cost, emissions, and efficiency trade-offs | |
GJO + PCGAN for multi-MGs with EVs | Lowers emissions, costs, and losses in EV charging | Restricted to EV control; lacks hybrid optimization–prediction | Proposes GrCRA + PCRTAM-Net synergy for broader grid applications | |
Two-stage reactive power optimization | Enhances voltage stability in distribution networks | Focused only on reactive power; not multi-objective | Extends to multi-objective cost–emission–efficiency scheduling | |
MOEA/D-SF (Multiobjective Evolutionary Algorithm with Superiority of Feasible Solutions) | MOEA/D-SF (Multiobjective Evolutionary Algorithm with Superiority of Feasible Solutions) | MOEA/D-SF (Multiobjective Evolutionary Algorithm with Superiority of Feasible Solutions) | Integrate GrCRA (deep learning–based feature extraction and prediction) to enhance both convergence and diversity in Pareto solutions. | |
BiCo algorithm for DG placement + reconfiguration | Reduces cost, losses, and voltage deviations | No hybrid storage; no forecasting | Integrates Battery + SC with prediction | |
Multi-objective multiverse optimization (MVO) with parallel processing in bi-level framework (voltage → cost) | Multi-objective multiverse optimization (MVO) with parallel processing in bi-level framework (voltage → cost) | Multi-objective multiverse optimization (MVO) with parallel processing in bi-level framework (voltage → cost) | Multi-objective multiverse optimization (MVO) with parallel processing in bi-level framework (voltage → cost) | |
Advanced optimization for renewable grids | Integrates AI + evolutionary computation for uncertainty handling | Emphasizes optimization but lacks a hybrid predictive–control framework | Introduces a joint optimization–prediction approach for RES grids |
