Abstract: A method for computing possibly optimal policies in reinforcement learning with multiple objectives and tradeoffs includes receiving a dataset comprising state, action, and reward information for objectives in a multiple objective environment. Tradeoff information indicating that a first vector comprising first values of the objectives in the multiple objective environment is preferred to a second vector comprising second values of the objectives in the multiple objective environment is received. A set of possibly optimal policies for the multiple objective environment is produced based on the dataset and the tradeoff information, where the set of possibly optimal policies indicates actions for an intelligent agent operating in the multiple objective environment to take.

Abstract: An embodiment establishes a knowledge base based at least in part on sensor data received from a network. The embodiment generates a predicted performance parameter for a designated entity of the network using a first machine learning algorithm. The embodiment compares the predicted performance parameter to an actual performance parameter and determines whether the actual performance parameter exceeds a threshold difference from the predicted performance parameter. The embodiment generates, responsive to determining that the threshold difference is exceeded, incentive data using a second machine learning algorithm, where the incentive data is representative of an action selected by the second machine learning algorithm using an iterative optimization process, and where the iterative optimization process comprises performing the action and determining that the actual performance parameter approaches the threshold value in response to the action.

Abstract: According to one embodiment, a method, computer system, and computer program product for reinforcement learning is provided. The present invention may include training, using an offline dataset, a plurality of diverse reward models, and creating a policy based on an output of the reward models and a robustness operator of the reward models.

Abstract: Embodiments for assessing energy in a thermal energy fluid transfer system in a cloud computing environment by a processor. Behavior of the thermal energy fluid transfer system, associated with a heating service, a cooling service, or a combination thereof, may be learned according to collected data to identify one or more energy usage events. An energy usage assessment operation may be performed using temperature signal disambiguation operations, with data collected over a selected time period by one or more non-intrusive Internet of Things (IoT) sensors located at one or more selected positions in the thermal energy fluid transfer system, to learn the system performance indicators, and when coupled with ingested expected policy behavior, identify one or more energy usage waste events according to the learned behavior in real time.

Abstract: A processor may identify a first task of a set of tasks. The processor may identify features of the first task. The processor may generate a reputation assessment for a first user related to the features of the first task. The processor may match the first user to the first task based on the reputation assessment.

Abstract: A computer-implemented method for automated policy decision making optimization is disclosed. The computer-implemented method includes creating a dataset from a tabular database, wherein the dataset includes one or more columns selected as state variables, a column selected as action variables, and a column selected as reward variables. The computer-implemented method further includes determining a candidate function approximator Q based on applying at least one state variable, one action variable, and one reward variable to a trained regression model. The computer-implemented method further includes learning a decision policy based on applying the candidate function approximator Q to a reinforcement learning algorithm. The computer-implemented method further includes determining, based on the learned decision policy, an expected reward.

Abstract: A method for forecasting time delays added to a scheduled start time and a scheduled end time of a task includes generating a stochastic model of the task and resources affecting the task, the stochastic model includes a reactionary delay component that is a function of previous task end times and a root cause delay component that is an independent random process at a specific time. The method further includes: calculating a probability distribution of time delays added to the scheduled start time as a combination of the reactionary delay component and the root cause delay component using the stochastic model to provide a probability distribution of start times; and calculating a probability distribution of time delays added to the scheduled end time as a combination of the reactionary delay component and the root cause delay component using the stochastic model to provide a probability distribution of end times.

Abstract: A method for forecasting time delays added to a scheduled start time and a scheduled end time of a task includes generating a stochastic model of the task and resources affecting the task, the stochastic model includes a reactionary delay component that is a function of previous task end times and a root cause delay component that is an independent random process at a specific time. The method further includes: calculating a probability distribution of time delays added to the scheduled start time as a combination of the reactionary delay component and the root cause delay component using the stochastic model to provide a probability distribution of start times; and calculating a probability distribution of time delays added to the scheduled end time as a combination of the reactionary delay component and the root cause delay component using the stochastic model to provide a probability distribution of end times.

Abstract: Embodiments for assessing energy usage efficiency in a fluid transfer pumping system in a cloud computing environment by a processor. A rate of temperature decay may be determined over a selected time period using a temperature signal collected by one or more non-intrusive Internet of Things (IoT) sensors located at one or more selected positions of a piping network in the fluid transfer pumping system so as to determine energy efficiency in the fluid transfer pumping system associated with a heating service, a cooling service, or combination thereof.

Abstract: Embodiments for implementing intelligent refrigeration state classification in an Internet of Things (IoT) computing environment by a processor. A signal from a single IoT sensor associated with a refrigeration system may be used to assist in automatically classifying refrigeration states according to a training phase and an operational phase.

Abstract: Embodiments for assessing energy in a thermal energy fluid transfer system in a cloud computing environment by a processor. Behavior of the thermal energy fluid transfer system, associated with a heating service, a cooling service, or a combination thereof, may be learned according to collected data to identify one or more energy usage events. An energy usage assessment operation may be performed using temperature signal disambiguation operations, with data collected over a selected time period by one or more non-intrusive Internet of Things (IoT) sensors located at one or more selected positions in the thermal energy fluid transfer system, to learn the system performance indicators, and when coupled with ingested expected policy behavior, identify one or more energy usage waste events according to the learned behavior in real time.

Abstract: Embodiments for intelligent flow prediction by a processor. One or more flows of a domain of interest between target entities may be forecasted according to one or more forecast models learned via machine learning using extracted features of one or more target variables from one or more data sources. The one or more flows may include a quantitative value, an intensity score, an intensity category, or a combination thereof between the target entities.

Abstract: Embodiments for assessing energy usage efficiency in a fluid transfer pumping system in a cloud computing environment by a processor. A rate of temperature decay may be determined over a selected time period using a temperature signal collected by one or more non-intrusive Internet of Things (IoT) sensors located at one or more selected positions of a piping network in the fluid transfer pumping system so as to determine energy efficiency in the fluid transfer pumping system associated with a heating service, a cooling service, or combination thereof.

Abstract: A method for forecasting time delays added to a scheduled start time and a scheduled end time of a task includes generating a stochastic model of the task and resources affecting the task, the stochastic model includes a reactionary delay component that is a function of previous task end times and a root cause delay component that is an independent random process at a specific time. The method further includes: calculating a probability distribution of time delays added to the scheduled start time as a combination of the reactionary delay component and the root cause delay component using the stochastic model to provide a probability distribution of start times; and calculating a probability distribution of time delays added to the scheduled end time as a combination of the reactionary delay component and the root cause delay component using the stochastic model to provide a probability distribution of end times.

Abstract: A method for forecasting time delays added to a scheduled start time and a scheduled end time of a task includes generating a stochastic model of the task and resources affecting the task, the stochastic model includes a reactionary delay component that is a function of previous task end times and a root cause delay component that is an independent random process at a specific time. The method further includes: calculating a probability distribution of time delays added to the scheduled start time as a combination of the reactionary delay component and the root cause delay component using the stochastic model to provide a probability distribution of start times; and calculating a probability distribution of time delays added to the scheduled end time as a combination of the reactionary delay component and the root cause delay component using the stochastic model to provide a probability distribution of end times.