Struct vrp_core::models::problem::GoalContext

pub struct GoalContext { /* private fields */ }

Defines Vehicle Routing Problem variant by global and local objectives: A global objective defines the way two VRP solutions are compared in order to select better one: for example, given the same amount of assigned jobs, prefer less tours used instead of total solution cost.

A local objective defines how single VRP solution is created/modified. It specifies hard constraints such as vehicle capacity, time windows, skills, etc. Also it defines soft constraints which are used to guide search in preferred by global objective direction: reduce amount of tours served, maximize total value of assigned jobs, etc.

Both, global and local objectives, are specified by individual features. In general, a Feature encapsulates a single VRP aspect, such as capacity constraint for job’ demand, time limitations for vehicles/jobs, etc.

Implementations

impl GoalContext

pub fn new(
    features: &[Feature],
    global_objective_map: &[Vec<String>],
    local_objective_map: &[Vec<String>]
) -> Result<Self, String>

Creates a new instance of VrpVariant with features specified using information about hierarchy of objectives.

impl GoalContext

pub fn accept_insertion(
    &self,
    solution_ctx: &mut SolutionContext,
    route_index: usize,
    job: &Job
)

Accepts job insertion.

pub fn accept_route_state(&self, route_ctx: &mut RouteContext)

Accepts route state.

pub fn accept_solution_state(&self, solution_ctx: &mut SolutionContext)

Accepts solution state.

pub fn merge(&self, source: Job, candidate: Job) -> Result<Job, ViolationCode>

Tries to merge two jobs taking into account common constraints. Returns a new job, if it is possible to merge them together having theoretically assignable job. Otherwise returns violation error code.

pub fn evaluate(&self, move_ctx: &MoveContext<'_>) -> Option<ConstraintViolation>

Evaluates feasibility of the refinement move.

pub fn estimate(&self, move_ctx: &MoveContext<'_>) -> Cost

Estimates insertion cost (penalty) of the refinement move.

Trait Implementations

impl Clone for GoalContext

fn clone(&self) -> GoalContext

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Default for GoalContext

fn default() -> GoalContext

Returns the “default value” for a type.

impl MultiObjective for GoalContext

type Solution = InsertionContext

The solution value type that we define the objective on.

fn total_order(&self, a: &Self::Solution, b: &Self::Solution) -> Ordering

An objective defines a total ordering between any two solution values.

fn fitness<'a>(
    &'a self,
    solution: &'a Self::Solution
) -> Box<dyn Iterator<Item = f64> + 'a>

An objective fitness values for given solution.

fn get_order(
    &self,
    a: &Self::Solution,
    b: &Self::Solution,
    idx: usize
) -> Result<Ordering, String>

Get solution order for individual objective.

fn get_distance(
    &self,
    a: &Self::Solution,
    b: &Self::Solution,
    idx: usize
) -> Result<f64, String>

Gets solution distance for individual objective.

fn size(&self) -> usize

Returns total number of inner objectives.

impl Shuffled for GoalContext

fn get_shuffled(&self, random: &(dyn Random + Send + Sync)) -> Self

Returns a new instance of ObjectiveCost with shuffled objectives.

impl HeuristicObjective for GoalContext