Abstract: Methods and systems for performing non-linear optimization of reach are described herein. The methods and systems may be used to compute the reach associated with different advertisement campaigns involving different combinations of spots. The method includes computing probabilities that a user will access a certain spot. The method includes retrieving a weight associated with the targeted user demographic. The method includes computing a first reach value using a first non-linear function, and a second reach value using a second non-linear function. The method includes comparing the two reach values to each other and to a predetermined reach threshold. The higher reach value larger than the predetermined reach threshold is selected and information associated with the selected reach value is provided to the user.

Abstract: Methods and systems for performing non-linear optimization of reach are described herein. The methods and systems may be used to compute the reach associated with different advertisement campaigns involving different combinations of spots. The method includes computing probabilities that a user will access a certain spot. The method includes retrieving a weight associated with the targeted user demographic. The method includes computing a first reach value using a first non-linear function, and a second reach value using a second non-linear function. The method includes comparing the two reach values to each other and to a predetermined reach threshold. The higher reach value larger than the predetermined reach threshold is selected and information associated with the selected reach value is provided to the user.

Abstract: Methods and systems are disclosed for using multiple reach optimization models together. These models may be used by an optimization system that helps an advertisement campaign manager decide how advertisements should be allocated between different media assets. In some embodiments, the models may be run simultaneously while their results and compared, and execution of some models may be ended based on the comparisons. In some embodiments, models may be combined together to create new models. In some embodiments, results from a first model may used as input in a second model running concurrently. In some embodiments, results from a previous execution of a model may be used to warm-start a subsequent execution of the model.

Abstract: Methods and systems are disclosed for using multiple reach optimization models together. These models may be used by an optimization system that helps an advertisement campaign manager decide how advertisements should be allocated between different media assets. In some embodiments, the models may be run simultaneously while their results and compared, and execution of some models may be ended based on the comparisons. In some embodiments, models may be combined together to create new models. In some embodiments, results from a first model may used as input in a second model running concurrently. In some embodiments, results from a previous execution of a model may be used to warm-start a subsequent execution of the model.

Abstract: Methods and systems are presented for creating and retrieving a matrix wherein a first dimension represents viewers, a second dimension represents media assets, and each data point represents a probability that a viewer accessed a media asset. The methods further include reducing the number of rows and columns of the matrix based on a desired size of the matrix and calculating a reach value for a media asset based on a the data points in the modified matrix.