Verteego Doc
  • Getting started
    • About Verteego
    • Sample Use Cases
    • Concepts
  • Data
    • Introduction
    • Datasources
      • URL connector specification
    • Datasets
  • Pipelines
    • Forecasting Pipelines
      • Getting started
      • Configuration
        • Identifying and preparing data
          • calculated_cols
          • cols_type
          • date_col
          • normalize
          • preprocessing
          • prediction_resolution
        • Configuring the Forecasting Algorithm
          • algo_name
          • algo_type
          • algorithm_parameters
          • fit_parameters
          • conditional_algorithm_parameters
        • Building the Training and Prediction Set
          • column_to_predict
          • features
          • input_prediction_columns
        • Using Hyperparameter Tuning for the Model
          • tuning_search_params
          • hyperparameter_tuning_parameters
        • Evaluating the Results of the Forecast
          • scores
        • Modifying the results of the forecast
          • postprocessing
      • Calculators
        • External source
          • get_from_dataset
          • weather
        • Mathematic
          • aggregate_val_group_by_key
          • binary_operation
          • count_rows_by_keys
          • hierarchical_aggregate
          • mathematical_expression
          • unary_operation
          • Moving Average (EWM)
        • Machine Learning
          • pca
          • clustering
          • glmm_encoder
          • one_hot_encode
          • words_similarity
        • Transformation
          • fillna
          • fill_series
          • case_na
          • interval_index
          • constant
          • cyclic
          • replace
        • Temporal
          • bank_holidays_countdown
          • bankholidays
          • date_attributes
          • date_weight
          • day_count
          • duration
          • events_countdown
          • seasonality
          • tsfresh
    • Optimization Pipelines
      • Getting started
      • Configuration
      • Constraints
        • Unary Constraint
        • Binary Constraint
        • Aggregation Constraint
        • Order Constraint
        • Multiplicative Equality Constraint
        • Generate constraints from a Dataset
  • Apps
    • About Apps
    • Recipes
      • Pipelines
      • Datasets
  • Users
    • User roles
  • Best practices
    • Performance analysis and ML model improvement
  • Developers
    • API
    • Change logs
Powered by GitBook
On this page
  • Types of Constraints Supported
  • Dynamic Constraint Generation from Datasets
  1. Pipelines
  2. Optimization Pipelines

Constraints

PreviousConfigurationNextUnary Constraint

Last updated 1 year ago

Verteego supports a versatile set of constraints that can be applied to optimize various business operations effectively. These constraints ensure that the solutions provided by the platform not only achieve the desired outcomes but also adhere to all specified operational limits and conditions.

Types of Constraints Supported

  • Description: A Unary Constraint applies a single condition to one variable within the dataset. This type of constraint is typically used to set limits (such as minimum or maximum values) on individual variables.

  • Use Case: Ensuring that the quantity of a product does not exceed warehouse capacity.

  • Description: Binary Constraints involve relationships between two variables. These are used to establish a direct comparison or interaction, such as equality or inequality.

  • Use Case: Linking sales volume to stock levels to avoid overproduction.

  • Description: This constraint deals with aggregated data across a set of variables. It is useful for conditions that involve sums, averages, or other aggregate calculations.

  • Use Case: Maintaining budget limits by capping the total expenditure across multiple departments.

  • Description: An Order Constraint imposes an order among the values of different variables, ensuring that they follow a specified sequence or hierarchy.

  • Use Case: Sequencing of project tasks where certain tasks must logically precede others.

  • Description: These constraints involve equalities that include multiplicative terms of variables. They are complex and used to model interactions where variables are dependent on the product of others.

  • Use Case: Adjusting resource allocation proportionally based on varying demand levels.

Dynamic Constraint Generation from Datasets

Verteego provides the capability to dynamically generate constraints based on real-time or historical data, enhancing the flexibility and applicability of the optimization process:

  • Functionality: Users can specify rules and conditions directly through the interface, choosing relevant columns and defining how these should translate into constraints.

  • Advantages: This feature allows constraints to be adaptive to changes in data, ensuring that the optimization process remains relevant and effective under varying business conditions and data states.

Unary Constraint
Binary Constraint
Aggregation Constraint
Order Constraint
Multiplicative Equality Constraint
Learn how to generate constraints from Datasets >>