Когда возникает неожиданная проблема со здоровьем, отдел неотложной помощи больницы (ED) становится нашим жизненно важным медицинским ресурсом. Повышение эффективности и своевременности ухода за ЭД при одновременном сокращении допустимых повторных госпитализаций чревато трудностями, связанными со сложностью и неопределенностью. В этой статье мы описываем новую систему поддержки принятия решений, сочетающую машинное обучение, моделирование и оптимизацию для решения этих задач. Система позволяет администраторам здравоохранения оптимизировать рабочий процесс во всем мире, принимая во внимание неопределенность в отношении возникающих травм и заболеваний и связанной с ними помощи, тем самым значительно сокращая продолжительность пребывания пациента. Это достигается без изменения физической структуры, а вместо этого путём фокусирования на консолидации процессов, отслеживании операций и укомплектовании персоналом. Впервые эта система была внедрена в больнице Грейди Мемориал в Атланте, штат Джорджия, что позволило сократить продолжительность пребывания пациента в б�

"We describe a software application that enables owners of generation output from a virtualized Federal Columbia River Power System to safely operate the system while also shaping the generation to meet their energy and economic needs. The application, known colloquially as The Optimizer, employs modern Operations Research techniques to convert a highly non-linear problem into a linear one so as to create a robust solution for the entire, six-dam system on an hourly basis; over a ten-day time horizon; within seconds to minutes. The tool permits two operators to manage and optimize the entire generation portfolios for nine utilities simultaneously in a very stringent time frame around the clock, and enables planners to ensure that the operation of the river meets all of the requirements for flood control, fish management, electrical reliability, safe dam operation, and recreation under high degrees of uncertainty. As a result, the Optimizer allows utilities to integrate renewable, environmentally friendly wind and solar generation into their resource portfolio with hydro generation and empowers fast decision making and adaptation to rapidly changing conditions."

"LMI разработал решение для управления запасами товаров с редким спросом (изолированные всплески спроса), а также изделиями с частым, но весьма изменчивым спросом.Прогнозирование спроса на эти товары - независимо от того, насколько сложным был метод прогнозирования, - привело к годам плохих результатов, например, к избыточному запасу некоторых товаров. В конечном итоге, агентство достигло своих целей, связанных с инвентаризацией, улучшением обслуживания клиентов и снижения рабочей нагрузки покупателей, при этом не увеличивая запасы - все это благодаря покупке большего количества того, что продается и меньше того, что нет". "LMI developed the PNG inventory control solution to manage items with infrequent demand (isolated spikes in demand) as well as items with frequent, but highly variable demand. Such items account for the majority of stocked hardware at the Defense Logistics Agency (DLA). Forecasting demands for these items-no matter how sophisticated the forecasting method-had led to years of bad outcomes for DLA: excess inventory for some items, backorders for others, and excessive buyer workload. The implementation of PNG (a software package of two inventory solutions, Peak Policy and Next Gen) shifted DLA from trying to forecast individual items to a portfolio or risk management approach to inventory control. PNG presents a three-way tradeoff between customer service, inventory value, and replenishment workload. DLA can then make a single decision t

"We partnered with the online flash sales retailer Rue La La to develop and implement a pricing decision support tool that sets initial prices for new products. Rue La La is in the online fashion sample sales industry, where they offer extremely limited-time discounts (typically 2-3 days) on designer apparel and accessories. Our approach is two-fold and begins with developing a demand prediction model for new products. The two biggest challenges faced when building our demand prediction model include estimating lost sales due to stockouts, and predicting demand for styles that have no historical sales data. We use descriptive analytics (clustering) and predictive analytics (regression) to address these challenges and predict future demand. Regression trees - an intuitive, yet non-parametric regression model - prove to be effective predictors of demand."

"Ingram Micro, the world's largest technology distributor operates in a high volume low margin environment. The company started its Business Intelligence & Analytics practice in North America (NA) about 6 years ago. Since then the group has built and deployed a scalable highly innovative price optimization engine, imprime™, for NA's $8B spot business, a set of analytics apps, imsmart™ for its internal sales organization and an integrated digital marketing platform, intelligence.ingrammicro.com, to run data-driven marketing campaigns to its customers and end customers. Since 2011, these products and analytical programs have yielded a cumulative benefit of $1.12 Billion of incremental product revenue and $28M of incremental gross profit. These solutions have been effective in driving profitable growth and are scalable. Our next steps are to continue to drive these best practices within the other regions outside NA and continue our sales enablement activities."

"The planning of spare engines and engine parts is a challenging and important task for airlines to seamlessly support flying and engine repair operations. Engines are expensive and critical assets that make this problem important from the financial and operational perspectives. In this talk, we present an application of a simulation-based approach utilized for planning the ownership level of spare engines and engine parts required to achieve a target service level or out-of-service (OTS) aircraft performance metrics. Our application considers both single and multi-location settings as well as two levels of modeling: a higher level where the repairable components are the engines as a whole, and a lower level that focuses on the repair of the individual engine parts. "

"The Mayo Clinic Department of Orthopedic Surgery was facing low utilization of their operating rooms (ORs) for spine surgical procedures, combined with fluctuating empty days and days with overtime to complete scheduled surgeries. Extremely long days ended up being unsafe days with increased provider fatigue and higher likelihood of errors. Investigation revealed the cause to be inaccurate estimation of surgical and non-surgical duration and scheduling of surgeries rather than limited surgery demand. Existing scheduling optimization research in the literature was inadequate, as they provided a single "optimal" solution. Often, the single optimal solution was not satisfactory to the patients and the providers. The team 1) conducted descriptive research using historical data to identify clinical and operational factors; 2) developed and implemented predictive models for the duration of surgical and nonsurgical times in the OR based on these factors; and 3) developed and implemented a prescriptive scheduling search algorithm that suggests multiple slots for a given surgery thereby providing flexibility while ensuring high probability of completion of surgical day without much overtime (completion closer to 5 pm)."

"Long term infrastructure planning of locomotive service facilities is vital to the efficiency of the railroad. We developed a large-scale optimization model that integrates decisions on (i) location, capability, and capacity of fixed facilities, (ii) home location and routing plan of movable facilities, and (iii) assignment of a variety of service demands. A decomposition-based solution framework was developed and shown to bring significant economic benefits in full-scale implementations."

"We present the problem of scheduling crash tests for new vehicle programs at Ford. We developed a completely custom-made scheduling system that transforms a labor-intensive scheduling process relying on high levels of expertise, to a more automated one that utilizes optimization and institutionalizes expert knowledge. Our system enables engineers and managers to consider multiple scheduling scenarios, using efficient interfaces to specify problem instances and efficient methods to solve them."

Вебинар Д.А. Гаврилова - управление запасами в цепи поставок

Вебинар о преимуществах использования языка моделирования AMPL при разработке оптимизационных приложений

Видеозапись презентации Llamasoft о необходимости постоянной работы по моделированию и стратегической оптимизации ЦП

Решение задачи проектирования логистической сети дистрибьютора спортивных товаров с помощью Anylogic

Архив примеров применения инструментов моделирования ЦП LLamasoft

Архив видеозаписей презентаций финалистов конкурсов в сфере аналитики, проводимых INFORMS - Wagner и Franz Edelman Award

Доклад С. Крола об имитационном моделировании службы Service Desk в банке с использованием Anylogic. Наиболее интересные моменты доклада - анализ статистических данных для формирования распределений входных параметров модели, а также решение многокритериальной задачи оптимизации с использованием функции желательности Харрингтона

Что делать, чтобы записать формализацию математической (оптимизационной) модели?