Table 1 Comparative analysis of representative approaches.

Model Predictive Control -based works⁵⁰

PV ± storage

Model Predictive Control⁵⁷

Simulation; metrics: tracking error, constraint satisfaction

Excellent stability and constraint handling³⁴

Requires an accurate model; limited multi-objective optimization^32,33

Combine machine learning forecasting with a hybrid optimiser to maintain constraint satisfaction and improve multi-objective trade-offs.

Genetic Algorithm (GA)/PSO/Differential Evolution (DE) optimizers^51,52,53

PV + battery/PV + wind

Metaheuristics (Genetic Algorithm, PSO, Differential Evolution)⁵⁸

Simulation: cost, emissions, efficiency

Good global search, multi-objective optimization³⁵

Premature convergence, parameter sensitivity, and high runtime^36,37

Hybrid global + local search, adaptive parameters, and parallelism reduce convergence and runtime issues.

Reinforcement Learning -based energy management system⁵⁴

PV + battery + demand response

Reinforcement Learning⁵⁹

Simulation; reward, cost

Learns model-free policies³⁸

High data demand, unstable training, weak guarantees^39,40

Use machine learning forecasting and constrained optimisation for safety and deterministic guarantees; Reinforcement Learning is reserved for future extension.

Forecasting + rule-based energy management system⁵⁵

PV + wind

Statistical forecasting with heuristics⁵⁸

Simulation/pilot; cost savings

Simple, robust, easy to implement^41,42

Suboptimal, inflexible to changes⁴³

Replace heuristics with adaptive optimization informed by IoT real-time data.

Pilot IoT deployment⁵⁶

PV + storage

Commercial energy management system with vendor rules⁵⁹

Small-scale pilots; uptime, user acceptance

Real-world proof-of-concept^44,45

Proprietary, not generalizable^46,47

The proposed framework proposes a scalable, open architecture with security and communication considerations.

Proposed System

PV + wind + battery + demand response (IoT integrated)

Machine learning forecasting + hybrid optimization (global heuristics + local refinement)

Smart-home simulation: system efficiency, energy cost, CO₂ emissions

Holistic IoT integration, multi-objective optimization, real-time adaptability⁴⁸

Requires pilot validation for scalability⁴⁹

Explicit IoT-driven framework, adaptive optimizer, quantified economic & environmental benefits.