It was from this perspective that the seminar by Prof. Giuseppe C. Calafiore, co-organized by the Center for AI Research (CAIR) and the Research Management Office (RMO) on the afternoon of April 14, 2026, raised a thought-provoking question: does optimizing charging cost always go hand in hand with optimizing emissions?

In his talk, “A Multi-Criterion Approach to Smart EV Charging: Coupling Dispatch-Informed Carbon Signals with Tractable EV Scheduling,” Prof. Giuseppe presented a quantitative framework for smart EV charging in which electricity cost and carbon emissions are treated as two objectives to be balanced rather than considered separately. The point of departure was a very clear reality: in power systems still significantly dependent on fossil fuels, the cheapest electricity hours are not necessarily the cleanest ones. On a representative day inspired by the Vietnamese context, cleaner hours tend to appear in late morning and early afternoon, when solar generation helps lower the grid’s carbon intensity, while the lowest-price hours are mostly overnight. This misalignment is precisely the central tension of the problem.

From there, the seminar guided the audience through an approach that is compact yet rich in practical significance. The research team builds a simplified dispatch layer to estimate hourly carbon intensity based on the active generation mix, including coal, gas, hydropower, photovoltaic, wind, and imports. This hourly carbon signal is then combined with the time-of-use electricity tariff to form the basis for an EV charging scheduling model that can be solved very quickly. What is particularly notable is that the problem does not change the total amount of energy that vehicles need to receive during the day; it only reshapes the timing of delivery within each vehicle’s feasible window. In other words, the optimization is not about charging “more” or “less,” but about when to charge in order to achieve a better balance between cost and emissions.

One of the most interesting parts of the seminar lay in the comparison among different strategies. When only electricity cost is optimized, the system tends to shift charging load toward low-price hours. By contrast, when the carbon signal is included in the objective, charging demand is progressively redirected toward cleaner time windows. Importantly, the results show that this trade-off does not necessarily imply sacrifice: in some settings, the system can maintain the same cost as the pure cost-optimal strategy while reducing charging-associated emissions; in other settings, a more carbon-focused strategy still performs better than the conventional first-in, first-out operation in both cost and CO₂ emissions. This is a particularly suggestive result, because it shows that sustainability is not always a matter of rigid trade-offs; sometimes it is a matter of redesigning the signals and decision mechanisms.

At a deeper level, the seminar was not only about EVs. It in fact touched on a broader issue in modern AI and optimization systems: when a local decision is embedded in a complex system, its effectiveness cannot be evaluated through a single metric alone. A charging hub does not control the national generation mix, yet it can still make a meaningful difference if it responds correctly to the signals emitted by the system. It is precisely here that a technical problem becomes a systems-thinking problem: if we want to build smarter operational technologies, we need models that are simple enough to run in practice, yet rich enough to capture the real tensions among cost, environment, and deployability.

The seminar highlighted a clearer way to look at real-world EV charging: when to charge is not only a cost decision, but also one that directly affects the carbon intensity of the power system at that moment. From this point of view, “smart charging” is no longer simply a matter of minimizing the electricity bill; it becomes part of a broader way of thinking about energy operation that is both more efficient and more sustainable.

 Explore CAIR’s research directions, ongoing projects, and opportunities to collaborate: https://cair.vinuni.edu.vn/