摘要:
To perform forecasting, a first data collection having data values at first intervals is received, and a first forecasting model is built based on the first data collection. A second forecasting model is built based on a second data collection having intervals aggregated from intervals of the first data collection, wherein the second forecasting model is at a different aggregation level than the first forecasting model. At least one metric is computed by performing at least one test based on at least one of the first and second data collections to indicate which of the first and second forecasting models has a better forecast quality.
摘要:
To perform forecasting, a first data collection having data values at first intervals is received, and a first forecasting model is built based on the first data collection. A second forecasting model is built based on a second data collection having intervals aggregated from intervals of the first data collection, wherein the second forecasting model is at a different aggregation level than the first forecasting model. At least one metric is computed by performing at least one test based on at least one of the first and second data collections to indicate which of the first and second forecasting models has a better forecast quality.
摘要:
One embodiment is a method that receives historical data of suppliers and applies, to the historical data, a mathematical optimization system that includes a set of mathematical equations and inequalities that express capabilities and capacities of the suppliers. The mathematical optimization system includes an objective function that minimizes a number of the suppliers to perform third-party labor services for an enterprise. The method selects a sub-set of the suppliers to perform the third-party labor services for the enterprise.
摘要:
A method for determining call center resource allocation can include modeling call center performance over an operations time period using a computer. A number of replicas of the modeled call center performance are simulated, using the computer, over a planning time period, each replica having random contact arrivals and contact service times following a stochastic arrival and service process according to a probability distributions of inter-arrival time and service time. Multiple iterations of each simulation are run on the computer to optimize call center resource allocation. A particular simulation iteration is tested against a criterion of convergence, and call center resource is allocated based on the particular simulation iteration with a successful criterion of convergence.
摘要:
A method for determining call center resource allocation can include modeling call center performance over an operations time period using a computer. A number of replicas of the modeled call center performance are simulated, using the computer, over a planning time period, each replica having random contact arrivals and contact service times following a stochastic arrival and service process according to a probability distributions of inter-arrival time and service time. Multiple iterations of each simulation are run on the computer to optimize call center resource allocation. A particular simulation iteration is tested against a criterion of convergence, and call center resource is allocated based on the particular simulation iteration with a successful criterion of convergence.
摘要:
The present disclosure is directed to product warranties having a residual value and a potential rebate based on the residual value. The residual value can thus be inversely related to the number of claims filed against the warranty.
摘要:
Methods, systems, and computer-readable and executable instructions are provided for determining a product price. Determining a product price can include determining an initial market attraction value, a market price sensitivity, and cost information for a product. Determining a product price can also include receiving a market constraint with respect to the product and pricing the product based on the initial market attraction value, the market price sensitivity, the cost information, and the market constraint.
摘要:
A system and method for determining the optimum price that a service provider should charge to customers of at least a partially refundable extended-product warranty to optimize profits generated from providing such warranties. In one aspect of the present invention the customer may elect to purchase warranty coverage when the product is new and cancel warranty coverage at any time thereafter, which election is based in part on the customer's expected discounted net utility from his coverage decisions. In another aspect of the present invention, the customer is allowed to make dynamic repair or replacement decisions in each period based on the product's failure status or on other criteria. In one embodiment, the customer can be afforded warranty coverage flexibility in terms of his ability to turn coverage off whenever desired and to obtain a partial refund of the warranty premium. By properly modeling extended-product warranty strategies from the perspective of the customer and from the perspective of the service provider, one can compute the customers' expected discounted net utility and the provider's expected discounted profit from strategic customers. In another aspect of the present invention a computer-based service is provided to the customer of the extended-product warranty for determining the customer's optimal dynamic decisions to maximize the customer's expected discounted net utility when making product replacement decisions, maintenance decisions, and warranty coverage decisions.
摘要:
A maximum expected value of a residual value warranty for a product to a customer is determined. An expected cost to a provider to support the residual value warranty for the customer is determined, based on the maximum expected value of the candidate residual value warranty to the customer. The expected profitability of the candidate residual value warranty is determined based on the expected cost.
摘要:
Parent node data is split into first and second child nodes based on a first partition variable to create a tree-based model. A first regression model for the first child node data relates the response variable and the predictor variable.