Optimization Service

AI-generated Key Takeaways

The Linear Optimization Service enables the modeling and resolution of linear and mixed-integer linear programs within Apps Script.
It provides classes like LinearOptimizationConstraint, LinearOptimizationEngine, and LinearOptimizationSolution to define, solve, and retrieve optimization results.
LinearOptimizationEngine allows adding variables, constraints, setting objective functions (maximization or minimization), and solving the linear program.
Solutions can be evaluated using methods like getObjectiveValue, getStatus, and getVariableValue to understand the optimization outcome.
The service utilizes various statuses (e.g., OPTIMAL, FEASIBLE, INFEASIBLE) and variable types (INTEGER, CONTINUOUS) to represent the solution state and variable characteristics.

The linear optimization service, used to model and solve linear and mixed-integer linear programs.

Name	Brief description
`LinearOptimizationConstraint`	Object storing a linear constraint of the form `lowerBound ≤ Sum(a(i) x(i)) ≤ upperBound` where `lowerBound` and `upperBound` are constants, `a(i)` are constant coefficients and `x(i)` are variables (unknowns).
`LinearOptimizationEngine`	The engine used to model and solve a linear program.
`LinearOptimizationService`	The linear optimization service, used to model and solve linear and mixed-integer linear programs.
`LinearOptimizationSolution`	The solution of a linear program.
`Status`	Status of the solution.
`VariableType`	Type of variables created by the engine.

Method	Return type	Brief description
`setCoefficient(variableName, coefficient)`	`LinearOptimizationConstraint`	Sets the coefficient of a variable in the constraint.

Method	Return type	Brief description
`addConstraint(lowerBound, upperBound)`	`LinearOptimizationConstraint`	Adds a new linear constraint in the model.
`addConstraints(lowerBounds, upperBounds, variableNames, coefficients)`	`LinearOptimizationEngine`	Adds constraints in batch to the model.
`addVariable(name, lowerBound, upperBound)`	`LinearOptimizationEngine`	Adds a new continuous variable to the model.
`addVariable(name, lowerBound, upperBound, type)`	`LinearOptimizationEngine`	Adds a new variable to the model.
`addVariable(name, lowerBound, upperBound, type, objectiveCoefficient)`	`LinearOptimizationEngine`	Adds a new variable to the model.
`addVariables(names, lowerBounds, upperBounds, types, objectiveCoefficients)`	`LinearOptimizationEngine`	Adds variables in batch to the model.
`setMaximization()`	`LinearOptimizationEngine`	Sets the optimization direction to maximizing the linear objective function.
`setMinimization()`	`LinearOptimizationEngine`	Sets the optimization direction to minimizing the linear objective function.
`setObjectiveCoefficient(variableName, coefficient)`	`LinearOptimizationEngine`	Sets the coefficient of a variable in the linear objective function.
`solve()`	`LinearOptimizationSolution`	Solves the current linear program with the default deadline of 30 seconds.
`solve(seconds)`	`LinearOptimizationSolution`	Solves the current linear program.

Property	Type	Description
`Status`	`Status`	Status of the solver.
`VariableType`	`VariableType`	Type of variables created by the solver.

Method	Return type	Brief description
`createEngine()`	`LinearOptimizationEngine`	Creates an engine to to solve linear programs (potentially mixed-integer programs).

Method	Return type	Brief description
`getObjectiveValue()`	`Number`	Gets the value of the objective function in the current solution.
`getStatus()`	`Status`	Gets the status of the solution.
`getVariableValue(variableName)`	`Number`	Gets the value of a variable in the solution created by the last call to `LinearOptimizationEngine.solve()`.
`isValid()`	`Boolean`	Determines whether the solution is either feasible or optimal.

Property	Type	Description
`OPTIMAL`	`Enum`	Status when an optimal solution has been found.
`FEASIBLE`	`Enum`	Status when a feasible (not necessarily optimal) solution has been found.
`INFEASIBLE`	`Enum`	Status when the current model is unfeasible (has no solution).
`UNBOUNDED`	`Enum`	Status when the current model is unbound.
`ABNORMAL`	`Enum`	Status when it failed to find a solution for unexpected reasons.
`MODEL_INVALID`	`Enum`	Status when the model is invalid.
`NOT_SOLVED`	`Enum`	Status when `LinearOptimizationEngine.solve()` has not been called yet.

Property	Type	Description
`INTEGER`	`Enum`	Type of variable that can only take integer values.
`CONTINUOUS`	`Enum`	Type of variable that can take any real value.