An off-grid system with renewable energy integration, considering a solar photovoltaic system, is a promising solution to fulfil the energy demand of consumers in remote areas. However, photovoltaic systems have uncertainties and variability issues which interrupt the regular operation and need proper planning for energy resilience. To handle the uncertainty problem of photovoltaic power, power-to-gas (PtG) technology can be integrated into the system, acting as long-term energy storage. To mitigate the challenges of uncertainties and variabilities, long-term energy storage, and resilient and economic operation, this research paper proposes a novel mathematical model of an integrated energy system with PtG technology for minimising the total system planning cost. A robust optimisation approach with the Karush-Kuhn-Tucker theorem is used to deal with bi-level optimisation and worst-case scenarios due to photovoltaic system uncertainty using PtG technology and energy storage system. Sensitivity analysis is performed with robust optimisation for proving the increase and decrease in the total cost. The results show an optimised result after applying robust optimisation to deal with photovoltaic system uncertainty, which increased the cost. The total cost, which includes investment and operation cost, is Rs. 20,554,932 (Gamma = 0) in the first year. As the uncertainty increased from 0 to 1, 2, & mldr;, 24, the cost also increased and at Gamma = 24, which is the worst case, the cost increased to Rs. 50,310,770.4. The final cost at year 19 is Rs. 12,010,000.11, which shows the minimised cost. The proposed approach decreases the overall cost by 2%-41.57% for 20 years' planning, which indicates economical and resilient operation in the future. The advantages of the proposed approach are justified by applying it to an off-grid system supplying energy with solar and gas equipment in a remote area, which can provide energy to 46-59 household consumers, depending on daily consumption (5.5-7 kWh/house).

The traditional passive energy consumption model does not adequately address the distribution network security constraints, and fair benefit distribution required for prosumers with distributed renewable energy and adjustable loads in the distribution grid. To address these gaps, this paper proposes an optimal energy transaction strategy to enhance distribution grid security while incentivizing prosumers. A bi-level Stackelberg game optimization model is developed between a virtual power plant (VPP) and an energy-sharing alliance. At the upper level, the VPP sets transaction prices to guide energy trading while ensuring grid security. At the lower level, prosumers form a peer-to-peer energy-sharing alliance, responding to VPP price signals. The model is transformed into a two-stage optimization problem using generalized Nash bargaining theory. The existence of a globally optimal solution is proven, and the bisection method is employed for solving the upper-level problem. The consensus alternating direction multiplier method is used for decentralized peer-to-peer transactions, preserving privacy and independence. By reformulating the model with Karush–Kuhn–Tucker conditions, a fully decentralized iterative solution is achieved. Simulations on the IEEE-33 node distribution grid confirm that the proposed model converges in six iterations. Multi-scenario analysis further validates the effectiveness of the benefit distribution mechanism, which accounts for prosumer contributions in the energy-sharing alliance.

The growing penetration of renewable energy sources and the electrification of transportation have introduced significant challenges in power system operations, including renewable intermittency, forecast uncertainties and increased peak demand. This paper presents an electric vehicle-integrated chance-constrained economic dispatch (EV-Integrated CCED) model, a novel framework that integrates electric vehicles (EVs) as distributed, bidirectional energy storage resources to address these issues. Unlike traditional models, the proposed approach incorporates a distributionally robust optimisation framework to handle uncertainties in renewable generation and net load forecasts, ensuring reliable and cost-efficient operation even under worst-case scenarios. By dynamically scheduling EV charging and discharging activities, the model enhances grid flexibility, optimises renewable energy utilisation and minimises operational costs. Numerical studies on the 8-zone ISO-NE test system demonstrate the model's ability to significantly outperform traditional methods, showcasing its potential for modern power systems transitioning to a clean and electrified energy future.

Virtual Power Plants (VPPs) serve as critical enablers for aggregating large-scale distributed resources to participate in spot and ancillary service markets, thereby enhancing power system flexibility and supply-demand coordination. However, the economics and operational efficiency of VPPs are constrained by market compatibility and the misalignment of bilateral willingness among VPPs and prosumers. This paper proposes a resource aggregation strategy that incorporates Mental Accounting Theory (MAT) into a bilateral matching framework to address two challenges: premium resource selection and matching stability. A differentiated bargaining-based bilevel decision-making model is formulated, wherein the upper level represents VPPs' portfolio optimization across multiple market segments, and the lower level captures the prosumers' individualized bargaining behaviors. A MAT-based boundary rationality description is proposed to reflect psychological effects such as loss aversion and reference dependence in VPPs-Prosumers bargaining process. Furthermore, a synergistic solution strategy based on Analytical Target Cascading method is developed for the bi-level model to derive bilateral preference rankings between VPPs and prosumers. Subsequently, a Gale-Shapley-based stable matching algorithm is architected to generate VPPs-prosumers pairings that satisfy individual rationality and global stability. Case studies on typical operational days demonstrate that the proposed method increases total VPPs revenue by 47.9% compared to conventional centralized planning, while ensuring stable and mutually acceptable matchings.

With the profound transformation of the global energy structure and the continuous advancement of the dual-carbon goal, virtual power plants (VPPs), with distributed energy as their core, have become a key technological path to improving grid flexibility, promote the consumption of new energy sources, and ensure energy security. However, the diversity of resources within VPPs, their wide geographical distribution, and the rapidly changing external market environment have brought unprecedented challenges to their operation and control. Automatic decision-making technology, especially advanced algorithms based on artificial intelligence (AI) and machine learning (ML), is becoming a core driving force for solving this problem. Therefore, this review introduces research on automatic decision-making technology in the operation and control of VPPs. First, it introduces the basic concepts, classifications, and operation and control framework of VPPs. Then, it describes the application methods of automatic decision-making technology in VPP operation and control. Finally, it points out the challenges of improving decision reliability and ensuring data security in VPP deployment and application. This provides theoretical support and practical guidance for subsequent research in the field of VPPs.

Local energy markets can operate in a centralized or distributed/decentralized way, which may lead to privacy leakage, computational burdens, or network violations. To address these issues, this paper proposes a standardized energy-trading method based on the equivalent projection (EP) theory, enabling rapid clearing of network-constrained local markets while preserving customer privacy. The bids/offers of prosumers, including operating objective and constraints, are transformed into standardized transaction-feasible regions (STFRs) with a preset error tolerance. The STFRs are mathematically represented by linear convex-hull inequality constraints via a progressive vertex enumeration (PVE) algorithm, which captures market efficiency. Case studies demonstrate the superiority of the proposed method.

Existing studies on shared energy storage have mainly focused on trading mechanisms, resource allocation, and pricing, while giving limited attention to price-quantity offering decisions that simultaneously account for resource coupling, state of charge, and market-price variations when aggregated VPP resources participate in multi-product shared-storage transactions. This study develops a multi-product price-quantity offering strategy for a VPP in a local shared energy storage market within an industrial park featuring coordinated operation of generation, the distribution network, loads, and storage. First, an existing double-auction mechanism was adapted to the operating characteristics of aggregated VPP resources. The tradable storage capability was divided into charging-power rights, discharging-power rights, and energy-capacity rights, and a multi-product double-auction mechanism was established between the VPP and storage-service users. Second, the coupled use of virtual-battery power and energy margins by the three products was represented through operating constraints and tradable limits. A period-specific price–quantity offering strategy was then formulated by combining market-reference-price forecasting, risk adjustment, and state-of-charge-dependent transaction costs. Subsequent offers were updated on a rolling basis using observed transaction prices. Finally, a market-clearing model was formulated to maximize bid-based social welfare and to allocate and settle shared storage resources between the VPP and individual users. Under the trading scenario and parameter settings considered in the case study, the proposed strategy dynamically adjusts its offers in response to resource states and market conditions. Compared with the maximum-quantity bidding benchmark, it yields higher net revenue and greater bid-based social welfare.

The construction of spot electricity markets plays a pivotal role in power system reforms, where market clearing systems profoundly influence market efficiency and security. Current clearing systems predominantly adopt a single-system architecture, with research focusing primarily on accelerating solution algorithms through techniques such as high-efficiency parallel solvers and staggered decomposition of mixed-integer programming models. Notably absent are systematic studies evaluating the adaptability of primary-backup clearing systems in contingency scenarios—a critical gap given redundant systems’ expanding applications in operational environments. This paper proposes a comprehensive evaluation framework for analyzing dual-system adaptability, demonstrated through an in-depth case study of the Inner Mongolia power market. First, we establish the innovative “Dual-Active Heterogeneous” architecture that enables independent parallelized operation and fault-isolated redundancy. Subsequently, key performance indices are quantitatively evaluated across four critical dimensions: unit commitment decisions, generator output constraints, transmission section congestion patterns, and clearing price formation mechanisms. An integrated fuzzy evaluation methodology incorporating grey relational analysis is employed for objective indicator weighting, enabling systematic quantification of system superiority under specific grid operating states. Empirical results based on actual operational data from 200 generation units demonstrate the framework’s efficacy in guiding optimal system selection, with particularly strong performance observed during peak load periods. The proposed approach shows high generalization potential for other regional markets employing redundant clearing mechanisms—particularly those with increasing renewable penetration and associated uncertainty.

The growing number of electric vehicles (EVs) provides virtual power plants (VPPs) with significant opportunities to deliver frequency regulation services; however, existing control strategies often treat batteries as disposable energy buffers, accelerating degradation and reducing their second-life economic value. Moreover, coordinating large-scale, heterogeneous EV fleets under strict real-time latency constraints remains computationally challenging for centralized approaches. This paper proposes a novel Smart-MFG framework that integrates lifecycle-aware cost modeling with a hierarchical multi-timescale control strategy for sustainable EV-grid interaction. A key contribution is the introduction of a state-dependent, lifecycle-aware shadow price that replaces conventional static degradation coefficients and serves as a unified coordination signal between the day-ahead bidding layer and the real-time dispatch layer. Specifically, a decoupled stochastic linear programming model performs day-ahead bidding under user uncertainty, while a clustering-based mean field game (MFG) controller governs fast real-time dispatch and supply-demand equilibrium among heterogeneous EV groups. To address scalability, the fleet is granulated using Rough Set Theory, preserving heterogeneity while reducing computational complexity. Case studies based on PJM regulation data with 1,000 EVs demonstrate that the proposed framework achieves high tracking accuracy (<0.8%) with low computational latency (<4 s), while maintaining competitive market profits (∼$1.55 M/year) and significantly improving lifecycle economic value.

This paper addresses the integration challenges of Distributed Energy Resources (DERs) within Virtual Power Plants (VPPs) arising from the confluence of stochastic human-centric behavior and physical variability. We propose a Cyber-Physical-Social System (CPSS) framework for VPP operation that employs a Priority-Driven Adaptive Robust Optimization (PD-ARO) strategy. Central to this framework is the System Urgency Index, a composite metric that quantifies real-time social risk, enabling the adaptive modulation of the system's uncertainty budget. To mitigate parameter drift and bridge the model-reality gap, an Urgency-Coupled Quantum Genetic Algorithm (UCQGA) is implemented for real-time cyber-layer calibration. Simulation studies on a modified IEEE 33-bus system demonstrate that the CPSS-based framework provides robust active power regulation and multi-energy coordination, maintaining nodal power balance and eliminating a worst-case 1.42 MW supply-demand shortfall under adversarial scenarios. Compared to deterministic and static robust benchmarks, the proposed model reduces real-time balancing costs by up to 81.9% while significantly enhancing operational reliability. The results confirm that explicit coordination of the cyber, physical, and social layers is a viable pathway for improving the economic efficiency and resilience of next-generation VPPs.

The increasing penetration of renewable-dominated entities such as virtual power plants (VPPs) amplifies electricity market uncertainty, driving demand for external information in trading decisions. Consequently, information markets, where information is treated as a tradable asset, becomes an inevitable development of modern electricity markets. However, the value of information (VoI) is difficult to quantify, and its quality is unverifiable prior to trading, leading to credibility deficiency, adverse selection, and market failure in information markets. Thus, this paper proposes a VoI-driven framework of credible information trading and pricing in VPPs. Specifically, a dynamic multi-dimensional credibility quantification model is proposed by introducing the concept of decision-effective quality, which endogenously incorporates information quality and credibility-related factors into decision utility. Meanwhile, a VoI-driven dynamic pricing mechanism is proposed to unify ex-ante pricing and ex-post verification of information trading. Furthermore, a three-level game-theoretic model of information provider (IfP), information demander (IfD) and trading platform (ITP) is formulated, along with a distributed solution algorithm integrating physics-informed neural networks (PINNs) with Anderson acceleration to enhance computational efficiency. Numerical studies based on real data from Italy electricity market demonstrate that compare to full-trust conditions, the proposed credibility mechanism avoids a 36.2% economic loss, and improves buyer profits by 6.1% compared to pure electricity market without information trading. The VoI-driven dynamic pricing boosts seller profit margins from 30.0% under fixed pricing to 73.0% while maintaining an 86.31% return on investment for buyers. The proposed algorithm achieves millisecond-to-second solution times in large-scale settings, exhibiting good scalability.

The rapid integration of distributed renewable energy resources has significantly increased voltage regulation challenges in distribution networks. Virtual power plant (VPP) technology is a promising solution for mitigating voltage fluctuations by leveraging the flexibility of demand-side resources. Among demand-side resources, heating, ventilation, and airconditioning (HVAC) systems possess substantial regulation potential. However, coordinating the HVAC systems within the VPP gives rise to a high-dimensional and nonlinear control problem, which is induced by HVAC system's characteristics (e.g., response delay and strong thermal coupling). In this context, this paper adopts an attention-enhanced factored multi-agent reinforcement learning method to address the above challenges. A value decomposition network is enhanced to capture the nonlinear relationship between HVAC operation characteristics and system states through an attention mechanism. In addition, a centralized policy gradient based on the factored value function is employed to optimize the joint action space to achieve better agent coordination. Case studies indicate that the proposed method can effectively mitigate voltage fluctuations. Meanwhile, the occupants' desired temperature is maintained while controlling HVAC systems.

With the implementation of Chinese “dual carbon” policy, virtual power plants (VPPs) have become a key instrument in modern power systems for integrating distributed renewable energy, enhancing grid flexibility, and promoting low-carbon development. However, the operational efficiency of a VPP is highly dependent on optimal resource allocation during the planning phase, particularly in electricity–carbon coupled markets where economic and environmental objectives must be balanced. To this end, this paper proposes a novel market-oriented VPP resource allocation strategy that incorporates electricity–carbon coupling mechanisms. A bi-level framework is established, in which the upper level determines the VPP resource allocation strategy and the lower level simulates prosumer operations. The resulting model is solved by an improved Analytical Target Cascading (ATC) algorithm that innovatively introduces adaptive penalty factors to enhance convergence. Case studies demonstrate the feasibility and superiority of the proposed strategy in achieving synergistic optimization of economic benefits and low-carbon objectives.

With the government's cancellation of subsidies for newly registered centralized photovoltaic (PV) power stations and the exacerbation of solar curtailment in China, PV and energy storage (PV-ES) investors are urgently in need of transitioning to market-oriented operational models to enhance their revenues. Distribution system markets and distributed transactions offer PV-ES investors channels for trading and opportunities for value enhancement. However, given the diverse personalized preferences and privacy protection requirements of prosumers within the local market, PV-ES investors lack a transactional decision-making methodology across multiple distribution network local energy markets (LEMs), making it challenging to discern the profit signals from different distribution network LEMs. This paper proposes a metamodel-based optimization algorithm for PV-ES investors to participate in multi-LEM transactions. A representative industrial park's local energy trading market is modeled, integrating electricity and carbon credit trading. A bilevel optimization model is then established to determine PV-ES investors' trading strategies across multiple LEMs within the park. To protect privacy, the LEM model is solved using the alternating direction method of multipliers (ADMMs). To address high computational demands, a hybrid solution algorithm combining differential evolution (DE) and dynamic partial least squares Kriging metamodel (HA-DEDKM) is proposed. The results show that the proposed strategy effectively enhances the profitability of PV-ES investors. The employed solution method avoids frequent invocation of lower-level market transaction models, significantly reducing computational load while preserving privacy and improving solution efficiency.

This paper proposes a bi-level optimal scheduling method for battery swapping stations based on equity profile stratification and dynamic incentives. First, an Equity Type-Replenishment Label system classifies users into three core groups. Second, a two-way dynamic incentive mechanism guides free-equity users toward residential charging via non-linear psychological expectations, while attracting paid users during off-peak periods. Based on this, a multi-station rolling horizon joint optimization model maximizes operator revenue. Simulations using real operational data show this mechanism accurately identifies user response thresholds, significantly increasing comprehensive profits and smoothing nodal voltage fluctuations without compromising user experience.

With the increasing penetration of renewable energy and the rapid growth of data-center electricity demand, traditional transmission network planning based solely on physical power grids faces growing challenges in terms of congestion mitigation, renewable energy accommodation, and investment efficiency. This paper investigates a stochastic transmission network planning method from the perspective of a virtual power grid, in which geographically distributed data centers are modeled as demand-side virtual power plants with spatial load-shifting capability enabled by communication networks. By reallocating computing workloads across regions, data centers can effectively reshape the spatial distribution of electricity demand without relying on physical transmission capacity. The proposed model accounts for uncertainties in conventional loads, data-center workloads, wind power output, and equipment availability, and aims to minimize the total cost consisting of transmission investment, system operation, and data-center workload scheduling. To address the large-scale mixed-integer nature of the problem and protect information privacy, a Benders decomposition-based solution approach is adopted, enabling coordinated optimization between the power grid and data centers. Case studies based on the IEEE 24-bus reliability test system demonstrate that the proposed method can significantly reduce transmission expansion costs, alleviate congestion, and enhance renewable energy utilization. The results further indicate that, under high renewable penetration and deep digitalization, incorporating virtual power grid capabilities into transmission planning provides an effective pathway toward a more flexible, economical, and resilient power system.

With the rapid growth of flexible loads in power systems, virtual power plants (VPPs) have emerged as a key technological platform for aggregating distributed resources to participate in demand response (DR) programs. The accuracy of aggregated baseline load (ABL) prediction is critical to the effectiveness of DR programs and the fairness of market incentives. However, this remains challenging since the ABL cannot be directly measured and differs from the actual load during DR programs. Additionally, the varying aggregation scales of VPPs create a heterogeneous customer base, further complicating ABL prediction. To address these challenges, this study proposes a demand response adaptive decomposition and prediction (DRADP) system for different types of power customers with varied load levels. Specifically, we first perform data extension and mode decomposition for various resource types, reconstructing ABL data into trend and fluctuation branches. Upon the completion of data construction, a model incorporating a patch temporal convolutional network, an encoder–decoder structure, and a multi-source attention mechanism is developed to capture hidden features and volatile trends. The forecasting results are then integrated and regressed into point and interval predictions. Finally, experiments based on real customer load data from VPP operators validate the effectiveness of the proposed system in both deterministic forecasting and uncertainty analysis. In representative 60-min DR forecasting scenarios, DRADP achieves the mean absolute percentage error values of MAPEDM=1.534%, MAPEBAC=2.520%, and MAPECBSF=5.862%, respectively, significantly outperforming the benchmark models.

The ongoing decline in system inertia during the energy transition poses serious challenges to frequency stability. Virtual power plants (VPPs) offer a feasible pathway to enhance system frequency support by aggregating distributed inertia resources from the demand side. To this end, this paper proposes an integrated framework that combines bidding and clearing mechanisms to incentivize VPPs to exploit their potential inertia resources. First, an inertia assessment model is developed to characterize the dynamic characteristics of heterogeneous devices within the VPP. On this basis, a bidding strategy incorporating prospect theory is designed, which introduces competitors’ bids as a reference to improve the VPP’s clearing probability while optimizing its expected revenue. Furthermore, a market-clearing model based on the system marginal pricing principle is established, aiming to minimize total social costs and provide effective price incentives for market participants. Case simulation results demonstrate that compared with a VPP using a conventional expected utility criterion, the proposed strategy significantly improves the VPP’s market clearance rate and increases the VPP’s total revenue by 870.31 CNY, representing a 17.5% relative improvement over the conventional expected utility–based bidding strategy. Meanwhile, the marginal pricing mechanism effectively reflects system inertia scarcity signals, helping to incentivize more inertia resources to participate in the market and ensuring operational security of the system.

The microgrid (MG) formation is widely utilized to restore loads for enhancing the distribution system (DS) resilience against disasters. However, the prevailing literature overlooks the persistent impact of lasting ice storms on MG, hindering the survival of lines and restored loads. To address this problem, this paper proposes a MG formation methodology that proactively adjusts the de-icing current to melt ice accretion on lines. Considering that the de-icing current is highly dependent on MG formation decisions, there is a prevailing decision-dependent uncertainty (DDU) that links equipment outages to deicing decisions. Accordingly, a DDU ambiguity set is proposed for outages to introduce MG formation into de-icing strategies. Then, a two-stage decision-dependent distributionally robust (DRO) model is established, where the first stage schedules the MG formation and de-icing operations, and the second stage determines the load restoration. Additionally, a customized column-and-constraint generation (C&CG) algorithm is designed to handle the non-convexity and non-linearity of the model derived from the binary recourse and the variable coupling in the DDU ambiguity set. This methodology is validated on the modified IEEE 33-node and 123-node test feeders. It demonstrates that the proposed MG formation method is effective in enhancing the resilience of DS against the unfolding ice storm.