A clear understanding of plastic waste collection potential in a locality is key to optimize the recycling system in terms of collection and disposal schedules, vehicle deployment, and personnel arrangements. This study aims to fill this gap by predicting individual collection demands of industrial plastic waste (IPW) using an artificial intelligence technique, demonstrated for multiple types of facilities in Fukuoka Prefecture, Japan.

We developed a machine learning approach to predict the amount of IPW to be collected from individual facilities based on daily waste manifest data. Through the explorations from multiple aspects, the best models were selected and high prediction accuracies were achieved for 5 types of facilities: supermarket (84.6％), hospital (84.4％), logistics company (78.2％), food manufacturing company (82.5％), and building management company (81.1％). The root mean square error performed not so effectively in selecting the optimal prediction model. The IPW collection from hospital was the easiest to predict accurately. Climate conditions were not useful for the hospital and logistics company but others. Moreover, we identified the robust prediction periods for these facilities from September 1, 2020: supermarket (20 days), hospital (2 days), logistics company (6 days), food manufacturing company (1 day), and building management company (2 days).

Instead of statistical approaches, artificial intelligence (AI) techniques have been utilized for waste management in many fields owing to their higher accuracy. It provides opportunities to make accurate future predictions of collection demands and detect the optimal collection routes. This study aims to address plastic waste management using AI by applying predicted individual collection demands of industrial plastic waste (IPW) to an integrated collection system, as demonstrated in the Fukuoka Prefecture, Japan.

We propose an AI-based approach for applying known collection demands of IPW regarding vehicle routing problems to better integrate the existing IPW collection system. After providing details on future prediction of the collection demands through the machine learning approach, the Euclidean-distance-optimized vehicle routing problem was solved using Python. To further validate this method, an optimal route was estimated for a real road network. Finally, reductions in traveling distance and carbon dioxide (CO₂) emissions were evaluated for the collection system both before and after AI-assisted integration.

In this study, a distance-optimized collection route was identified, thus demonstrating the feasibility of integrating existing collection systems using AI technology. This integration was proven to be beneficial in terms of the traveling distance (22 km reduced per collection, i.e., 14.2％ of the total distance was reduced) and CO₂ emissions (4.8 kg-CO₂ reduced per collection, i.e., 10.1％ of the total emissions were reduced).

This study focused on infectious wastes discharged from small-scale medical institutions and aimed to establish an IoT system to promote the proper treatment of medical wastes and improve small-scale collection operations’ efficiency. The present study investigated the proper treatment of infectious wastes in medical institutions, the actual discharge situation, collection, and transportation. Furthermore, we conducted on-site interviews and questionnaire surveys with three collection and transportation companies and 260 small medical institutions in I city. As a result, we developed and proposed a traceability system based on the field needs obtained from these surveys. Specifically, we introduced a button-type pickup request system that replaces the telephone pickup request system, as well as a traceability system based on the activity record management system, and an information-sharing system based on the cooperation of multiple collection and transport companies.

Improving of waste treatment efficiency is one of the most important issues for intermediate treatment facilities. The waste treatment planning in intermediate treatment facilities depends on the past experience and intuition of workers, which may not be the best decision. Moreover, the waste may not arrive as planned. Most previous studies have focused on waste collections, transportations, and facility locations. We previously proposed a linear programming-based waste treatment planning model in intermediate treatment facilities. The numerical results showed that the proposed model can reduce total cost comparing to the actual data and meet the treatment demands even if the wastes do not arrive as planned. However, this model may cause unnecessary operations of machines due to non-selective machine assignment. In this paper, we proposed a modified version of the previous waste treatment planning model to solve these problems. This model is formulated as a 0-1 integer programming problem. We also show the effectiveness of our proposed model by comparing its performance with the actual data.

In the production process of end of life vehicles (ELVs), skilled workers inspect the ELVs visually. The dependency on skilled workers is high, causing problems related to ageing, knowledge transfer, and quality variation. In this study, we investigated the possibility of improving the production process of recycled automobile parts by automating a part of the parts inspection process. Specifically, we investigated the possibility of introducing an AI-based image diagnosis system and developed and evaluated a prototype of the system. We conducted a preliminary study at a production plant of recycled automobile parts and found that the time required for the production and inspection process of doors was relatively large among the exterior parts. In addition, we found that “scratches” accounted for a large proportion of the damage to exterior parts. Therefore, we conducted an experimental study to determine the presence or absence of scratches on doors using AI. The results showed that the loss function decreased, and the accuracy of the AI system was about 97％ after 100 training sessions. As a result of considering the extension to the damage other than “scratches”, it was suggested that the method could be applied to other damage such as “dents”, “rust”, and “stepping stones” on the premise that the accurate teacher data of 500 to 1,000 pieces are secured. This means the system can cover about 65％ of the total damage. On the other hand, one of the challenges for this system’s practical use is establishing a method of collecting accurate supervisory data, and it is expected that the system will be upgraded on the assumption that AI will perform the image diagnosis.

To prevent climate change, the movement toward early realization of carbon neutral society is accelerating in Japan and in the world. If steam supply from waste incinerators to manufacturing plants becomes common, it would be effective from the viewpoint of efficient reduction of CO₂ emission from the industrial sector and efficient utilization of waste. However, implementation of such a system has been rare in Japan. Therefore, we have conducted a study aiming the examination and demonstration of a challenge to promote the commercialization of steam supply projects by sharing information with potential stakeholders of the projects. Then, we proposed a system for stable steam supply, which is a barrier to the commercialization of steam supply projects, and evaluated its feasibility in terms of economy. Out of the 40 institutions that were provided information and interviewed, 37 institutions showed high interest in commercialization, and 28 institutions were conducting or supporting feasibility studies for commercialization as of December 1st in 2021. Thus, it was found that information sharing would promote the efforts for commercialization. It was also shown that the mechanism of stable steam supply could be economically introduced.

Efforts are underway to apply ICT and AI for the advancement of social infrastructure maintenance. In the venous supply chain of industrial waste, the contribution of technology and equipment is particularly significant in the incineration of industrial waste. While there is not enough room for facilities, equipment and human resources, it is necessary to increase the utilization rate of the facilities and maintain a certain level of technology and quality over a long period of time. In this study, we conducted a demonstration test of preventive maintenance monitoring of blowers and fabric filter cloths at industrial waste incineration plants by using information and communication technologies, and also tried to estimate the effect of maintenance cost reduction.

In recent years, information and communication technology has become popular in many industries, but it is expected to be introduced into safety management in particular. In this study, we introduced a system to measure the biological information of workers with different work in a waste treatment facility and evaluated the actual situation on the labor intensity using smart wear. As a result, we confirmed the case of dangerous labor intensity where the same work, but different the labor intensity for each worker. It was suggested that these results depend on the characteristics of each individual and may not be dealt with by unified safety management.

The efficiency improvement of inputting information on paper manifest is an issue in the office work of industrial waste disposal companies. AI-OCR and RPA were introduced on a trial basis in order to verify the possibility and cost-effectiveness of efficiency improvement and automation of paper manifest input work by information technology. We confirmed the usage status of multiple forms of paper manifests, grasped the rate of correct reading of input information by AI-OCR, and measured the work time of transfer to the core business system by RPA.

The ratio of paper manifest forms that were used at five or more per day (defined as the ‘standard form rate’) was as high as approximately 90％, and the percentage of correct answers to all items read using AI-OCR was 87.0％ in average. AI-OCR has a function to learn the reading result in the past, so the reading accuracy may be improved, if the number of registration increases in future. Especially, it is necessary to unify full-width and half-width and address notation in order to raise the reading accuracy. In addition, as for the transcription work time by RPA, the data input time (including confirmation and correction) per 1 sheet was 50.2 seconds/sheet for a typical input pattern. In the case of the atypical input pattern, the automation was difficult, and the conventional manual input was necessary. Based on this, the suggestion that it is realistic to divide the automation work and the non-automation work first, and to raise the proportion of the former which applies RPA was obtained.

The recycling rate has been decreasing in China because of the growing average income. The Chinese government has introduced a new recycling system based on the internet. Under the system the residents can be subsidized if they separate garbage into each material. We conduct a questionnaire to examine the effectiveness of Huge Recycling Application on recycling in Hangzhou. The method for empirical analysis in this article is based on Randomized Controlled Trial (RCT). We find that garbage separation application will stimulate recycling activities of residents. According to our empirical results, the recycling rate has been raised by 16–20％ since such an application was activated in Hangzhou.

As the impacts of climate change on agriculture are becoming more apparent, adopting immediate and effective adaptation measures to reduce climate change effects is essential. Despite the importance of adaptation, there is little evidence on studies about farmer adaptation and adoption factors. Hence, this paper aims to explore the characteristics and issues related to farmers’ adaptation and quantitatively examine the factors influencing adaptation measures. The main results of this study, based on a web questionnaire survey and meteorological data, are as follows. The questionnaire results show that more than half of the farmers adopt adaptation measures, of which the majority are adopted voluntarily. On the other hand, we found problems such as a limited number of farmers adopting multiple adaptations and maladaptation due to set-up costs and lack of information. The results of the econometric analysis on adaptation factors revealed that age, farm size, and perception of adverse effects of climate change are associated with the adoption of adaptation measures. Moreover, this study shows that risk and time preferences, as well as climate conditions, also affect adaptation. These results suggest that measures to reduce farmers’ risks, such as lowering set-up costs, can effectively promote adaptation.

Scientific evidence in chemical risk management is of key importance in triggering corresponding management. However, uncertainties in scientific evidences exist in some cases to varying extents. In this paper, we explore scientific uncertainties in detail, based on historical incidents. For the purpose of the analysis, we focused on historical experiences based on Minamata disease, Itai-itai disease, and the so-called Suginami disease incidents. The time series of three incidents were analyzed and characterized in terms of uncertainties of scientific evidences, disagreement on scientific evidences, and governmental actions taken. We found that the extent of uncertainties in causal relationships has apparent relation to the categories of governmental actions. In all three cases, limited management actions were taken when scientific uncertainties were relatively large, although management actions with enforcement were taken up when the uncertainties were small enough. Our observations suggest that substantial management actions are necessary even when scientific uncertainties are so large that current risk assessment schemes cannot be applied. We believe that an appropriate framework that integrates precautionary-based approach to scientific evidence-based risk assessment is important to manage scientific uncertainties appropriately.