Advanced Decision Support Tools Research Articles

Animal Feed Formulation-Connecting Technologies to Build a Resilient and Sustainable System.

The unprecedented challenges presented by the increase in global population have placed substantial demands on the livestock industry for human nutrition, necessitating heightened animal productivity and leading to an increased demand for natural resources to produce animal feed. Feed producers are leading the charge, consistently refining formulations to adapt to the evolving needs of livestock, driven in part by the cost of over 50% associated with feed production. This paper critically analyses the pressing issues within feed formulation, addressing the requirement for environmentally sustainable practices amidst the challenges of climate change. The exploration extends to how advanced decision support tools can enhance formulation techniques and profitability and contribute to environmental sustainability. Through an in-depth review of current feed formulation technologies, encompassing their applications and limitations, this study aims to enhance the existing knowledge base. Additionally, we examined future trends, highlighting the essential role of connecting technologies to establish a resilient and sustainable system. The emphasis is on the potential of formulation techniques to positively impact the environment and enhance the overall quality and performance of the animals. This paper provides actionable insights to improve animal production by examining feed formulation models and decision support tools. The anticipated outcome is a more informed and sustainable decision-making process, addressing the multifaceted challenges confronted by the livestock industry and making contributions to global efforts in climate change mitigation and environmental stewardship in animal production agriculture.

IVF Stimulation - personalized, optimized, and simplified using an advanced decision-support tool: A randomized trial

We assessed the effectiveness of a clinical decision support tool to reduce total cumulative follicle-stimulating hormone dosage used, to eliminate the need for ultrasound exams after day 5 of an individual superovulation cycle, and to improve the number of high-quality embryos obtained. The design we used was a Randomized Control Trial in a private fertility center. This study included 93 women aged 25–45 years undergoing IVF. 48 Test and 45 Control participants included normal and poor responders and patients with polycystic ovarian syndrome. A clinical decision support tool was used to forecast stimulatory hormone dosing for an individual cycle based on follicle size distribution on day 1 and day 5. Cumulative stimulatory hormone doses, oocytes retrieved, number of Mii oocytes, total embryos, high-quality embryos obtained during the cycle, and clinical pregnancy rates was the main outcome measures. Test participants required significantly lower cumulative FSH doses during superovulation cycles (average 1883 IU test, 2530 IU control, p <0.01), with significantly higher numbers of total embryos (average 5.4 test, 3.5 control, p<0.01), and high-quality embryos (average 3.1 test, 1.2 control, p<0.01). Test participants had higher Mii follicles, although the difference was not statistically significant. The clinical pregnancy rate reported was significantly higher for test participants than control participants (52% test, 26% control, p<0.05). The test group had more poor responders and PCOS patients than the control group. In conclusion, the clinical decision support tool eliminated the need for ultrasound exams after day 5, reduced the doses of stimulatory hormone required, yielded significantly higher numbers of high-quality embryos, and resulted in higher clinical pregnancy rates.

Incorporating sparse model machine learning in designing cultural heritage landscapes

Managing, protecting, and the evolutionary development of historical landscapes require robust frameworks and processes for forming datasets and advanced decision support tools. Despite the great potential, using pattern language, machine learning, and regenerative and generative design tools has yet to be adopted in historic landscape research due to the need for suitable training datasets. To address this theoretical and technical gap, this paper describes a three-step workflow, namely photogrammetry, feature extraction and discriminative feature analytics, to help facilitate the use of advanced ML tools for cultural heritage decision support. Sparse-Learning-Modelling (SLM) was used to help with feature extraction from small datasets. The developed tool was successfully tested on the 3D point cloud models of 13 heritage sites, and these could be replicated in other heritage sites with distinctive Cultural DNA worldwide. The findings of this research can extend the discourse of adopting advanced AI/digital tools in heritage landscape design.

87 An Augmented Reality Burn Management Application to Guide Care in Austere Environments

Abstract Introduction Novel medical technologies are necessary to support virtual health in low bandwidth environments. When tele-mentoring is not possible, advanced clinical decision support tools can deliver essential medical knowledge to assist novice clinicians. These tools can fill knowledge gaps in specialty fields, like burn. We developed the Augmented Reality Burn Assist Manager (ARBAM), a novel, comprehensive augmented reality (AR)-based burn management platform to address key tasks in burn care. Methods A randomized 2x2 cross-over design was used to investigate if ARBAM could enhance a user’s ability to care for a simulated burn patient as compared to standard clinical practice guidelines (CPGs). We developed the software for use on a mixed reality, head-mounted device. Participants were clinicians with little to no burn care experience, and were asked to complete 4 key tasks, twice each, using either paper copies of the Joint Trauma System’s CPGs and supplemental worksheets first (Paper) or our AR technology. Tasks included calculating burn size (TBSA), performing medication dosage calculations (MC), managing a burn fluid resuscitation (FR), and performing a simulated escharotomy (EC). All tasks were evaluated for completion time and accuracy. Cross-over analysis of treatment effects used a two-sided, two-sample t-test or Wilcoxon Rank-Sum test as appropriate. Results Eleven participants were randomized: 5 started with Paper and 6 started with AR. Time to perform FR was statistically different (p< 0.05) with AR taking less time than Paper (4.8 vs. 7.0 mins, respectively). FR clinical performance was also better in the AR group with a mean accuracy score of 99% vs. 52% for Paper (p< 0.001). EC completion times were not clinically or statistically different. However, EC accuracy was 93% for AR compared to 55% for Paper (p< 0.001). Although not statistically different, MC performance accuracy was considerably better for AR (91%) than for Paper (66%). No differences between groups were observed for TBSA or MC completion times nor TBSA accuracy. Conclusions Compared to traditional tools, AR may enhance a clinician's ability to complete key tasks in the initial management of a burn patient. This technology may be helpful during disasters when tele-mentoring is not available. Limitations of this study include a small sample size, limited user training on the AR device, and potential carry-over effects. Therefore, generalizations cannot be made about wide-scale usage. Nonetheless, this study showed that AR may be helpful in managing lifesaving burn tasks. Applicability of Research to Practice ARBAM may enhance decision making by novice clinicians when caring for burn patients. Our data demonstrate the potential for AR-based support tools to improve burn care delivery and reduce error rates for novice clinicians, which could lead to reductions in morbidity and mortality. Future research is needed to investigate how AR technology can be used for other lifesaving interventions.

Minimization of drug interactions in polypharmacy treatments of diabetes mellitus type 2 with cardiovascular comorbidities by using the decision support tool PM-TOM

BackgroundCombined polypharmacy treatments of multi-diseases like diabetes mellitus type 2 (DMT2) with its comorbidities could lead to serious adverse reactions (ADR) due to drug-drug interactions (DDIs). This study aimed to demonstrate that these DDI ADRs can be significantly reduced by carefully examining DDIs of recommended drugs and using advanced clinical decision support (CDS) tools, like PM-TOM (Personal Medicine: Therapy Optimization Method). MethodDMT2 with heart failure (HF) and atherosclerotic cardiovascular disease (ASCVD) were taken for analysis. First, 20 drug classes were selected, recommended in relevant medical guidelines (US, European and Canadian); for example, biguanides, sodium-glucose transporter 2 inhibitors, glucagon-like peptide-1 receptor agonists, insulins, angiotensin 2 receptor blockers, angiotensin-converting enzyme inhibitors, beta-adrenergic blockers, diuretics, and statins. Next, these classes were combined into polypharmacy treatment cases, which were organized into three groups: Basic (combinations of three drug classes), Medial (five), and Advanced (eight). Then, the tool PM-TOM was used to find treatments with minimal and maximal drug interactions (MIN-DDI and MAX-DDI) for each case. Finally, these two treatments' minimal, average and maximal DDIs were calculated and statistically analyzed to examine the scope and effects of optimizing polypharmacy treatments in each case group. ResultsIn the Basic group, 16 polypharmacy treatment cases were created; in the Medial 210 and the Advanced 736. The MIN-DDI and MAX-DDI treatments in each case group showed significant DDI differences; for example, in the Basic group, the average DDI count in the MIN-DDI treatments was 0.19 and in the MAX-DDI ones 1.75, while in the Medial and Advanced groups, these indicators were 1.66 and 7.66, and 5.76 and 20.52, respectively. Also, 87% of optimized treatments (MIN-DDI) in the Basic group showed no DDIs, 37% in the Medial, and 9% in the Advanced. In addition, 70% of cases in the Medial group had at most two DDIs, and 49% in the Advanced group at most five. ConclusionsThese findings suggest that DDI ADRs in randomly selected (unoptimized) DMT2 polypharmacy treatments can be substantially reduced using specialized decision support tools, increasing patients' chances for successful treatment and decreasing health care costs. Similar findings can be expected for other multi-diseases as well.

Towards the coordinated and fit-for-purpose deployment of Unmanned Aerial Systems (UASs) for flood risk management in England

Abstract Preparedness for flood emergency response is crucial for effective flood management. The need for advanced flood decision support tools that aid flood management has been recognized by several authors. This work examines the variability that currently exists across England with regard to the Unmanned Aerial System (UAS) data collection and processing strategy in flood emergency events. Expert elicitation was carried out using a tailored questionnaire about UAS deployment in three flood emergency scenarios. The survey highlighted that reduced equipment assembly time, a national network of appropriately qualified UAS pilots and the effective UAS deployment when on-site, can reduce the response time to flood emergency. For improved comparability and reduced bias in data collection and interpretation, clear guidelines on which data products are most beneficial for particular purposes, processing time required, platform and sensor selection may also be necessary. We consider that releasing a comprehensive documentation pack, which includes guidelines, standards and protocols that detail the methods, tools, technology, quantity and quality of data, to UAS pilots on a flood emergency call, will enhance the timely response.

Decision support system for adaptive sourcing and inventory management in small- and medium-sized enterprises

• Proposes a data-driven approach in sourcing and inventory management for SMEs. • Exploits Observational and empirical data to generate decision support solutions. • Promote the adoption and initiation of artificial intelligence in SMEs. • Promising implementation results in a real-world case study. Elevated business uncertainties and competition over recent years have caused changes to the data-driven supply chain management of sourcing and inventories across industries. However, only large-sized enterprises have the resources to harness data for aiding their decision-making and planning. By contrast, small- and medium-sized enterprises (SMEs) commonly have limited resources and knowledge, which affects their ability to collect and utilize data. Thus, it is a challenge for them to implement advanced decision support tools to mitigate the effects of market uncertainties. This paper proposes a decision support system (DSS) for sourcing and inventory management, with the aims of helping SMEs compile and exploit data, and supporting their decisions under business ambiguities. The DSS was developed using a simulation-optimization approach by incorporating an artificial neural network and a genetic algorithm for problem representation and optimizing decision support solutions. The exploitation of observational and empirical data reduces the burden of data compilation obtained from unorganized data sources across SME operations. Further, uncertainty factors such as raw material demand, price, and supply lead time were considered. When implemented in a medium-sized food industry company, the DSS can provide decision support solutions that integrate the selection of recommended suppliers and optimal order quantities. It can also help decision-makers to shape their inventory management policies under uncertain raw material demands, while also considering service levels, sales promotions, lead times, and material availability from multiple suppliers. Consequently, implementation of the DSS helped to reduce the total purchased raw material costs by an average of 51.62% and reduced the order interval and on-hand inventory costs by an average of 54.24%.

Artificial intelligence, nano-technology and genomic medicine: The future of anaesthesia.

Nanotechnology with artificial intelligence (AI) can metamorphose medicine to an extent that has never been achieved before. AI could be used in anesthesia to develop advanced clinical decision support tools based on machine learning, increasing efficiency, and accuracy. It is also potentially highly troublesome by creating insecurity among clinicians and allowing the transfer of expert domain knowledge to machines. Anesthesia is a complex medical specialty, and assuming AI can easily replace the expert as a clinically sound anesthetist is a very unrealistic expectation. This paper focuses on the association and opportunities for AI developments and deep learning with anesthesia. It reviews the current advances in AI tools and hardware technologies and outlines how these can be used in the field of anesthesia.

A multibranch, multitarget neural network for rapid point-source inversion in a microseismic environment: examples from the Hengill Geothermal Field, Iceland

SUMMARYDespite advanced seismological techniques, automatic source characterization for microseismic earthquakes remains difficult and challenging since current inversion and modelling of high-frequency signals are complex and time consuming. For real-time applications such as induced seismicity monitoring, the application of standard methods is often not fast enough for true complete real-time information on seismic sources. In this paper, we present an alternative approach based on recent advances in deep learning for rapid source-parameter estimation of microseismic earthquakes. The seismic inversion is represented in compact form by two convolutional neural networks, with individual feature extraction, and a fully connected neural network, for feature aggregation, to simultaneously obtain full moment tensor and spatial location of microseismic sources. Specifically, a multibranch neural network algorithm is trained to encapsulate the information about the relationship between seismic waveforms and underlying point-source mechanisms and locations. The learning-based model allows rapid inversion (within a fraction of second) once input data are available. A key advantage of the algorithm is that it can be trained using synthetic seismic data only, so it is directly applicable to scenarios where there are insufficient real data for training. Moreover, we find that the method is robust with respect to perturbations such as observational noise and data incompleteness (missing stations). We apply the new approach on synthesized and example recorded small magnitude (M ≤ 1.6) earthquakes at the Hellisheiði geothermal field in the Hengill area, Iceland. For the examined events, the model achieves excellent performance and shows very good agreement with the inverted solutions determined through standard methodology. In this study, we seek to demonstrate that this approach is viable for microseismicity real-time estimation of source parameters and can be integrated into advanced decision-support tools for controlling induced seismicity.

A Review of Current Applications of Artificial Intelligence and Machine Learning in Medical Science

Artificial Intelligence and Machine Learning are two fields that are causing substantial development in every field specifically in the field of medical sciences; for the stupendous potential that it can provide to assist the clinicians, researchers, in clinical decision making, automate time consuming procedures, medical imaging, and more. Most implementations of AI/ML rely on static data set, and this where the big data steps in. That is, these models are developed and trained on a data set that is already recorded and have been diligently reviewed for accuracy; leading to a precise decision-making process. Experts foresee that AI/ML based overarching care system will develop high-quality patient care and innovative research, aiding advanced decision support tools. In this paper we shall realize what are the current devices that are build and are being used for real time problem solving, also discuss the impact of Software as a Medical Device (SAMD) in future of medical sciences. [2,3,11]

Systems modelling and simulation to inform strategic decision making for suicide prevention in rural New South Wales (Australia).

The need to understand and respond to the unique characteristics and drivers of suicidal behaviour in rural areas has been enabled through the Australian Government's 2015 mental health reforms facilitating a move to an evidence-based, regional approach to suicide prevention. However, a key challenge has been the complex decision-making environment and lack of appropriate tools to facilitate the use of evidence, data and expert knowledge in a way that can inform contextually appropriate strategies that will deliver the greatest impact. This paper reports the co-development of an advanced decision support tool that enables regional decision makers to explore the likely impacts of their decisions before implementing them in the real world. A system dynamics model for the rural and remote population catchment of Western New South Wales was developed. The model was based on defined pathways to mental health care and suicidal behaviour and reproduced historic trends in the incidence of attempted suicide (self-harm hospitalisations) and suicide deaths in the region. A series of intervention scenarios were investigated to forecast their impact on suicidal behaviour over a 10-year period. Post-suicide attempt assertive aftercare was forecast to deliver the greatest impact, reducing the numbers of self-harm hospitalisations and suicide deaths by 5.65% (95% interval, 4.87-6.42%) and 5.45% (4.68-6.22%), respectively. Reductions were also projected for community support programs (self-harm hospitalisations: 2.83%, 95% interval 2.23-3.46%; suicide deaths: 4.38%, 95% interval 3.78-5.00%). Some scenarios produced unintuitive impacts or effect sizes that were significantly lower than what has been anticipated under the traditional evidence-based approach to suicide prevention and provide an opportunity for learning. Systems modelling and simulation offers significant potential for regional decision makers to better understand and respond to the unique characteristics and drivers of suicidal behaviour in their catchments and more effectively allocate limited health resources.

Bringing new tools, a regional focus, resource-sensitivity, local engagement and necessary discipline to mental health policy and planning

BackgroundWhile reducing the burden of mental and substance use disorders is a global challenge, it is played out locally. Mental disorders have early ages of onset, syndromal complexity and high individual variability in course and response to treatment. As most locally-delivered health systems do not account for this complexity in their design, implementation, scale or evaluation they often result in disappointing impacts.DiscussionIn this viewpoint, we contend that the absence of an appropriate predictive planning framework is one critical reason that countries fail to make substantial progress in mental health outcomes. Addressing this missing infrastructure is vital to guide and coordinate national and regional (local) investments, to ensure limited mental health resources are put to best use, and to strengthen health systems to achieve the mental health targets of the 2015 Sustainable Development Goals.Most broad national policies over-emphasize provision of single elements of care (e.g. medicines, individual psychological therapies) and assess their population-level impact through static, linear and program logic-based evaluation. More sophisticated decision analytic approaches that can account for complexity have long been successfully used in non-health sectors and are now emerging in mental health research and practice. We argue that utilization of advanced decision support tools such as systems modelling and simulation, is now required to bring a necessary discipline to new national and local investments in transforming mental health systems.ConclusionSystems modelling and simulation delivers an interactive decision analytic tool to test mental health reform and service planning scenarios in a safe environment before implementing them in the real world. The approach drives better decision-making and can inform the scale up of effective and contextually relevant strategies to reduce the burden of mental disorder and enhance the mental wealth of nations.

Global optimization of large-scale MIQCQPs via cluster decomposition: Application to short-term planning of an integrated refinery-petrochemical complex

Integrated refinery-petrochemical facilities are complex systems that require advanced decision-support tools for optimal short-term planning of their operations. The problem can be formulated as a mixed-integer quadratically constrained quadratic program (MIQCQP), in which discrete decisions select operating modes for the process units, while the entire process network is represented by input-output relationships based on bilinear expressions describing yields and stream properties, pooling equations, fuels blending indices and cost indicators. We develop a novel decomposition-based algorithm for deterministic global optimization that divides the network into small clusters according to their functionality. Inside each cluster, we derive a mixed-integer linear programming (MILP) relaxation based on piecewise McCormick envelopes, dynamically partitioning the variables that belong to the cluster and reducing their domains through optimality-based bound tightening. Results for an industrial case study in Colombia show profit improvements above 10% and significantly reduced optimality gaps compared with the state-of-the-art global optimization solvers BARON and ANTIGONE.

Artificial intelligence in dentistry: current applications and future perspectives

Artificial intelligence (AI) encompasses a broad spectrum of emerging technologies that continue to influence daily life. The evolution of AI makes the analysis of big data possible, which provides reliable information and improves the decision-making process. This article introduces the principles of AI and reviews the development of AI and how it is currently being used. AI technology has influenced the health care field because of the need for accurate diagnosis and superior patient care. In order to understand the trend of AI in dentistry, electronic searching was carried out, combined with approaching individual companies to obtain the details of AI-based services. The current applications of AI in clinical dentistry were introduced and summarized. In the future, the AI-based comprehensive care system is expected to establish high-quality patient care and innovative research and development, facilitating advanced decision support tools. The authors believe that an innovative inter-professional coordination among clinicians, researchers, and engineers will be the key to AI development in the field of dentistry. Despite the potential misinterpretations and the concern of patient privacy, AI will continue to connect with dentistry from a comprehensive perspective due to the need for precise treatment procedures and instant information exchange. Moreover, such developments will enable professionals to share health-related big data and deliver insights that improve patient care through hospitals, providers, researchers, and patients.

Seeking compromise across competing goals in conservation translocations: The case of the ‘extinct’ Floreana Island Galapagos giant tortoise

Abstract Conservation translocation projects must carefully balance multiple, potentially competing objectives (e.g. population viability, retention of genetic diversity, delivery of key ecological services) against conflicting stakeholder values and severe time and cost constraints. Advanced decision support tools would facilitate identifying practical solutions. We examined how to achieve compromise across competing objectives in conservation translocations via an examination of giant tortoises in the Galapagos Islands with ancestry from the extinct Floreana Island species (Chelonoidis niger). Efforts have begun to populate Floreana Island with tortoises genetically similar to its historical inhabitants while balancing three potentially competing objectives – restoring ecosystem services (sustaining a high tortoise population size), maximizing genome representation of the extinct C. niger species and maintaining a genetically diverse population – under realistic cost constraints. We developed a novel approach to this conservation decision problem by coupling an individual‐based simulation model with generalized additive models and global optimization. We identified several incompatibilities among programme objectives, with quasi‐optimal single‐objective solutions (sets of management actions) differing substantially in programme duration, translocation age, incubation temperature (determinant of sex ratio) and the number of individuals directly translocated from the source population. Quasi‐optimal single‐objective solutions were able to produce outcomes (i.e. population size and measures of genetic diversity and C. niger genome representation) to within 75% of their highest simulated outcomes (e.g. highest population size achieved across all simulations) within a cost constraint of c. $2m USD, but these solutions resulted in severe declines (up to 74% reduction) in outcomes for non‐focal objectives. However, when all programme objectives were equally weighted to produce a multi‐objective solution, all objectives were met to within 90% of the highest achievable mean values across all cost constraints. Synthesis and applications. Multi‐objective conservation translocations are likely to encounter complex trade‐offs and conflicts among programme objectives. Here, we developed a novel combination of modelling approaches to identify optimal management strategies. We found that solutions that simultaneously addressed multiple, competing objectives performed better than single‐objective solutions. Our model‐based decision support tool demonstrates that timely, cost‐effective solutions can be identified in cases where management objectives appear to be incompatible.

Medication decision-making for patients with renal insufficiency in inpatient and outpatient care at a US Veterans Affairs Medical Centre: a qualitative, cognitive task analysis

BackgroundMany studies identify factors that contribute to renal prescribing errors, but few examine how healthcare professionals (HCPs) detect and recover from an error or potential patient safety concern. Knowledge of...

Effect of smart meter data collection frequency in an early detection of shorter-duration voltage anomalies in smart grids

Smart grids are required to respond quickly and efficiently to any types of electrical incidents. This is possible only if advanced monitoring and decision support tools are available for the operators to collect and analyze the real-time data from the entire power system. Sensors and smart meters become necessary for an effective smart grid operation, as they can play a significant role in measuring system parameters such as the temperature of the transmission line, power outages and power usage. These sensors allow communication between the generation side and user side to ensure full network observability. Hence, the volume of data which needs to be analyzed for more reliable electrical services has been increased. Upon the analysis of all the data collected, useful insights can be gained to bring benefits to the utilities, consumers and the related organizations. In this project, the effect of smart meter data collection frequencies on the early detection of short-duration voltage anomalies has been investigated. An anomaly detection algorithm that analyzes the voltages collected from smart meters is developed. The proposed anomaly detector checks the normality of the voltage data collected from the smart meters installed at the residential loads. To test the effectiveness of the frequent-and-large smart meter data, a smart grid model was simulated in MATLAB/Simulink and the model was tested under three different operating conditions. The efficacy of the anomaly detector with the smart meter data collected at different data collection intervals was compared under the three operating conditions. The investigation of the smart meter data clearly shows that the shorter-duration anomalies can be detected effectively with more frequent data and hence it reveals the potential of smart meters in the early detection of short-duration anomalies in smart grids.

Future of Artificial Intelligence in Anesthetics and Pain Management

The potential of the second wave of Artificial Intelligence (AI) to change our lives beyond recognition is both exciting and challenging. AI has been around for over three decades, and this new approach of artificial intelligence, due to enhancements in technology, both software, and hardware, has resulted in the fact that human decision-making is considered inferior and erratic in many fields: none more so than medicine. Machine learning algorithms with access to large data sets can be trained to outperform clinicians in many respects. AI’s effectiveness in accurate diagnosis of various medical conditions and medical image interpretation is well documented. Modern AI technology has the potential to transform medicine to a level never seen before in terms of efficiency and accuracy; but is also potentially highly disruptive, creating insecurity and allowing the transfer of expert domain knowledge to machines. Anesthetics is a complex medical discipline and assuming AI can easily replace experienced and knowledgeable medical practitioners is a very unrealistic expectation. AI can be used in anesthetics to develop, in some respects, more advanced clinical decision support tools based on machine learning. This paper focuses on the complexity of both AI developments, deep learning, neural networks, etc. and opportunities of AI in anesthetics for the future. It will review current advances in AI tools and hardware technologies as well as outlining how these can be used in the field of anesthetics.

Data analytics-based decision support workflow for high-mix low-volume production systems

In order to answer the ever-fluctuating demand of high-mix low-volume production environments, reconfiguring the production systems and improving their performance rely heavily on the application of advanced decision support tools. Estimating the expected values of the performance measures (KPIs) in the face of these decisions, however, is even more challenging in such an environment as the complex structure, behavior and input demand creates an enormously large variable domain restraining the analysis. The paper introduces a novel workflow for providing simulation-based decision support for improving KPIs of high-mix low-volume production systems by reducing the size of the input domain with the application of unsupervised machine learning techniques.

Productivity and energy efficiency assessment of existing industrial gases facilities via data envelopment analysis and the Malmquist index

The current trend towards improving energy efficiency in industry calls for advanced decision-support tools for quantifying the level of efficiency of industrial facilities. This work applies Data Envelopment Analysis (DEA) to assess the performance of a set of 34 Air Separation Units (ASUs) producing industrial gases via air distillation. We identify the best ASUs according to energy efficiency and productivity criteria and define improvement targets for the units found inefficient. Furthermore, we analyze the temporal evolution of the efficiency scores using the Malmquist Productivity Index (MPI), which is calculated from real data of a gas company operating ASUs around the world. Our results provide insight on how to improve the efficiency of existing plants by identifying sources of inefficiency and reference facilities that could be used for benchmarking.