package operations_research.scheduling.jssp

Get desktop application:
View/edit binary Protocol Buffers messages

Stores how a job is executed.

Used in: JsspOutputSolution

• repeated AssignedTask tasks = 1

  How each task is executed.

• int64 due_date_cost = 2

  Earliness-Tardiness cost of that job.

• int64 sum_of_task_costs = 3

  Sum of all tasks costs for that job.

Stores how a task is executed.

Used in: AssignedJob

• int32 alternative_index = 1

  Indicates which alternative was selected. It corresponds to the alternative_index-th machine in the 'machine' field in Tasks

• int64 start_time = 2

  The start time of that task.

A job is an ordered sequence of tasks, plus hard constraints on its earliest start time, and its latest completion time. As well as optional earliness-tardiness penalties on its end date. The job starts with the first task in the list, and ends with the last.

Used in: JsspInputProblem

• repeated Task tasks = 1

  The ordered sequence of tasks.

• optional google.protobuf.Int64Value earliest_start = 2

  This date, if set, specifies a hard constraint on when the job can start.

• int64 early_due_date = 3

  This date specifies the earliest time the job should end. If this is set, then the earliness_cost_per_time_unit should be set too with a positive value.

• int64 late_due_date = 4

  This date specifies the latest time the job should end. If this is set, then the lateness_cost_per_time_unit should be set too with a positive value. If both early_due_date and late_due_date are set, then early_due_date <= late_due_date must hold.

• int64 earliness_cost_per_time_unit = 5

  The cost model is a convex function \ / \ / \________/ The slopes of this shape are specified on the left by the earliness cost, and on the right by the lateness cost. These penalty parts start at the early and late due dates. For one penalty part to be active, both the date and a positive cost must be defined. All costs must be positive.

• int64 lateness_cost_per_time_unit = 6

• optional google.protobuf.Int64Value latest_end = 7

  This date, if set, specifies a hard constraint on when the job can end. If both earliest_start and latest_end are specified, then earliest_start <= latest_end must hold.

• string name = 16

  Optional. A name for the job. This will only be used for logging purposes.

Specifies a precedence relation between jobs. It states: start(second_job) >= end(first_job) + min_delay.

Used in: JsspInputProblem

• int32 first_job_index = 1

• int32 second_job_index = 2

• int64 min_delay = 3

The input of a problem.

• repeated Job jobs = 1

• repeated Machine machines = 2

• repeated JobPrecedence precedences = 3

• int64 makespan_cost_per_time_unit = 4

• optional google.protobuf.DoubleValue scaling_factor = 5

  If set, the cost coefficients are multiplied by this factor. It is set to 1000 by the parser when reading early tardy taillard problems where the weight of a job is a floating point value.

• int32 seed = 24

  Sometimes, the academic data files contain extra information. We store it in the input problem message.

• string name = 16

  Optional: Name of the problem.

The output of solving a jobshop problem.

• repeated AssignedJob jobs = 1

  The solution for all jobs.

• int64 makespan_cost = 2

  The makespan cost of that solution.

• int64 total_cost = 3

  The total cost of that solution.

Used in: JsspInputProblem

• optional TransitionTimeMatrix transition_time_matrix = 1

  Optional transition time matrix for this machine.

• string name = 16

  Optional. A name for a machine. This will only be used for logging purposes.

This message specifies a task inside a job.

Used in: Job

• repeated int32 machine = 1

  The alternative machines that can perform that task. Only one must be selected. We store the index of the machine in the main JsspInputProblem.

• repeated int64 duration = 2

  The corresponding duration for the alternative ways of performing this task. This list must have the same size as the machine_id list.

• repeated int64 cost = 3

  An optional cost for selecting one alternative way of performing the task against another. This list must either be empty, or has the same size as the above two lists.

Stores the transition time matrix between jobs on a given machine. If the initial job has n jobs, then time[i * n + j] will indicate the minimum delay between the end of a task of a job i performed on this machine and the start of a task of job j performed on the same machine. Nothing can be executed on that machine during this delay.

Used in: Machine

• repeated int64 transition_time = 1