Multi-agent systems for water distribution
The Tisza River has been a source of livelihood for the people living in its region for centuries. In recent years, however, the river and its surroundings have been subjected to more and more serious pollution by agriculture, industry, and the activities of cities. Numerous technical interventions were carried out on the river (A Tisza-szabályozásról folyó vita közepén kapott szívrohamot Vásárhelyi Pál, 2021); however, floods and droughts still pose a threat to both wildlife and the economy.
There are two dams on the Hungarian section of the Tisza, their task is to store huge masses of water for later use, for example for irrigation, although nowadays they mostly perform a recreational task. The dams are opened and closed by human-controlled engines, according to the water level measured by the sensors. If, for example, a flood wave is expected, the need for pre-opening of the sluice gates can be established based on the experience of previous years and the most recent measurements.
Muranyi and Koncsos (2022) examined how water could be distributed in the event of a flood so that it is not wasted but recharges the amount of groundwater and atmospheric water. For this, in addition to the existing water reservoirs, up to 2.36 km3 of additional storage capacity is proposed, as indicated in Figure 1.
Figure 1: Potential water storage areas in the Great Hungarian Plain, in the Tisza Valley (Murányi, 2022)
One of the important aspects of the designation of these areas is that they had to be wetlands before and that they had to meet the purpose of storage through topographic features or possibly minor technical interventions. Another very important aspect is what kind of land use takes place in the designated storage area. Today, the main land use is cultivation, but in order to maintain biodiversity, grasslands, wetlands, pastures, and floodplain forests are recommended. The reservoirs are filled simultaneously with the rising water level of the river, and then the sluice gates are closed when the water reaches the permissible level in the reservoir, so the water flows do not damage the environment.
This solution can be further developed as a multi-agent system (MAS). A MAS includes agents that are individual computer systems that are able to perform actions and make decisions automatically to meet user-defined objectives, on the owner’s behalf and in its best interest. (Wooldridge and Jennings, 1995). These agents must communicate and negotiate with each other to achieve a common goal, even though their targets differ from other agents (Wooldridge, 2003). Sluice gates would also be placed between the reservoirs, and sensors capable of monitoring groundwater could be installed in the area of the reservoirs. The gates would be controlled by agents, in order to distribute the water level, the water would be directed to an area where the groundwater is too low, from reservoirs where the water level is too high or sufficient. These agents have an independent purpose, namely to keep the amount of groundwater optimal in their own area in an automated manner without human intervention. However, this requires them to communicate and negotiate with each other. The basis of the negotiation is the need to increase the level of groundwater in those reservoirs which have a higher need. Targeted subsurface water replenishment greatly reduces the risk of drought, evaporation, sudden floods, and inland excess water.
Agents could perform better by planning their next move based on the results of their past decisions: using a multi-agent reinforcement learning system or incorporating game theory. The latter enables agents to analyze the possible outcomes of their planned action in addition to the results of their experience and make their decisions based on them. Both developments would greatly increase the number of successful decisions based on reasoning.
Agents can also perform predictive maintenance using reinforcement learning, in which case the best technician can be assigned to a gate that needs maintenance (Ruiz Rodríguez et al., 2022), which can extend the lifespan of the system. Leveraging MAS to control water distribution would require less manual intervention, resulting in less human error, and making the system more efficient.
In the world, there are more and more serious problems with high water shortages (Water Scarcity | Threats | WWF, 2023) due to pollution emitted by agriculture, industry, and cities, and finally, global warming (Glied Viktor., 2009), so this solution can be a major step forward in replenishing groundwater and freshwater supplies. While the discussion is focused solely on the distribution of floods of the Tisza River through leveraging MAS, global groundwater management can preserve biodiversity and freshwater resources around the world. While this technological development has huge benefits, it is important to note that the investment is quite expensive and requires a lot of consultation with the people who currently farm or raise livestock in the areas of the designated reservoirs. Overall, we can see that MAS dramatically improves the method of targeted groundwater replenishment, which increases the amount of fresh water and biodiversity while reducing the risk of drought, evaporation, floods, and inland water excess.
References:
Glied Viktor. (2009). Vízkonfliktusok : küzdelem egy pohár vízért. Publikon.
Murányi, G. (2022) Scientific establishment of strategic options for alternative flood protection solutions View project. Available at: https://www.researchgate.net/publication/361925566.
Muranyi, G. and Koncsos, L. (2022) ‘Analysis of nature based flood management in the Tisza River Valley, Hungary’, Pollack Periodica, 17(3), pp. 83–88. Available at: https://doi.org/10.1556/606.2022.00456.
Ruiz Rodríguez, M.L. et al. (2022) ‘Multi-agent deep reinforcement learning based Predictive Maintenance on parallel machines’, Robotics and Computer-Integrated Manufacturing, 78. Available at: https://doi.org/10.1016/j.rcim.2022.102406.
A Tisza-szabályozásról folyó vita közepén kapott szívrohamot Vásárhelyi Pál (2021). Available at: https://mult-kor.hu/a-tisza-szabalyozasrol-folyo-vita-kzepen-kapott-szivrohamot-vasarhelyi-pal-20210408.
Water Scarcity | Threats | WWF (2023). Available at: https://www.worldwildlife.org/threats/water-scarcity?fbclid=IwAR2HxLe3nyd3HTkMWXhAsD3W-VwrJCmnH5P3-ZinyejCZF2ZVz-NO9PHtUM.
Wooldridge, M. (2003) ‘Introduction to MultiAgent Systems – Introduction’, pp. 1–13.
Wooldridge, M. and Jennings, N.R. (1995) ‘Intelligent agents: Theory and practice’, The Knowledge Engineering Review, 10(2), pp. 115–152. Available at: https://doi.org/10.1017/S0269888900008122.