u-schedule 0.2.2

Domain-agnostic scheduling framework: job-shop models, dispatching rules, GA encoding, constraint programming.
Documentation

u-schedule

Scheduling framework in Rust

Crates.io docs.rs CI License

Overview

u-schedule provides domain models, constraints, validation, dispatching rules, and a greedy scheduler for scheduling problems. It builds on u-metaheur for metaheuristic algorithms and u-numflow for mathematical primitives.

Modules

Module Description
models Domain types: Task, Activity, Resource, Schedule, Assignment, Calendar, Constraint, TransitionMatrix
validation Input integrity checks: duplicate IDs, DAG cycle detection, resource reference validation
dispatching Priority dispatching rules and rule engine
scheduler Greedy scheduler and KPI evaluation
ga GA-based scheduling with OSV/MAV dual-vector encoding
cp CP-based scheduling formulation

Dispatching Rules

Rule Description
SPT Shortest Processing Time
LPT Longest Processing Time
EDD Earliest Due Date
FIFO First In First Out
SLACK Minimum Slack Time
CR Critical Ratio
ATC Apparent Tardiness Cost
WSPT Weighted Shortest Processing Time
MWKR Most Work Remaining
LWKR Least Work Remaining
MOPNR Most Operations Remaining
PRIORITY Job Priority
RANDOM Random Selection

GA Encoding

The GA module uses dual-vector encoding for job-shop scheduling:

  • OSV (Operation Sequence Vector) — Permutation encoding that determines operation processing order
  • MAV (Machine Assignment Vector) — Integer vector that assigns each operation to a specific machine (for flexible job shops)

Quick Start

[dependencies]
u-schedule = { git = "https://github.com/iyulab/u-schedule" }

use u_schedule::models::{Task, Activity, Resource};
use u_schedule::validation::validate_input;
use u_schedule::dispatching::{DispatchingEngine, Rule};

// Define tasks with activities
let task = Task::new("T1")
    .with_activity(Activity::new("A1", 30_000)); // 30 seconds

let resource = Resource::new("R1", "Machine 1");

// Validate input
let errors = validate_input(&[task], &[resource]);
assert!(errors.is_empty());

Build & Test

cargo build
cargo test

Academic References

  • Pinedo (2016), Scheduling: Theory, Algorithms, and Systems
  • Brucker (2007), Scheduling Algorithms
  • Blazewicz et al. (2019), Handbook on Scheduling
  • Haupt (1989), A Survey of Priority Rule-Based Scheduling

Dependencies

  • u-metaheur — Metaheuristic algorithms (GA, SA, ALNS, CP)
  • u-numflow — Mathematical primitives (statistics, RNG)
  • serde 1.0 — Serialization
  • rand 0.9 — Random number generation

License

MIT License — see LICENSE.

WebAssembly / npm

Available as an npm package via wasm-pack.

npm install @iyulab/u-schedule

Quick Start

import init, { run_schedule, solve_jobshop } from '@iyulab/u-schedule';

await init();
const result = run_schedule({
  jobs: [
    { id: "A", processing_time: 5.0, due_date: 10.0 },
    { id: "B", processing_time: 3.0, due_date: 8.0 },
  ],
  config: { rule: "EDD" }
});

Functions

run_schedule(input) -> ScheduleOutput

Priority dispatching on a flat job list (single or parallel machines). Supports 12 rules: SPT, LPT, EDD, FCFS, CR, WSPT, MST, S/RO, ATC, LWKR, MWKR, PRIORITY.

Input:

{
  "jobs": [
    { "id": "A", "processing_time": 5.0, "due_date": 10.0, "release_time": 0.0, "weight": 1.0 }
  ],
  "config": { "rule": "SPT", "num_machines": 1, "atc_k": 2.0 }
}

Output:

{
  "schedule": [{ "id": "A", "start": 0.0, "end": 5.0, "tardiness": 0.0, "machine": 0 }],
  "makespan": 5.0,
  "total_tardiness": 0.0,
  "machine_utilization": [{ "machine": 0, "busy_time": 5.0, "utilization": 1.0 }]
}

Times are in seconds. machine_utilization is present only when num_machines > 1.

solve_jobshop(input) -> JobShopOutput

GA-based job-shop scheduling with multi-machine routing and precedence constraints.

Input:

{
  "jobs": [
    {
      "id": "J1",
      "operations": [
        { "machine": "M1", "processing_time": 3.0 },
        { "machines": ["M2", "M3"], "processing_time": 2.0 }
      ],
      "due_date": 15.0
    }
  ],
  "num_machines": 3,
  "ga_config": {
    "population_size": 100, "max_generations": 200,
    "mutation_rate": 0.1, "seed": 42,
    "tardiness_weight": 0.5,
    "crossover": "POX", "mutation": "Swap"
  }
}

Output:

{
  "schedule": [{ "job_id": "J1", "operation": 1, "machine": "M1", "start": 0.0, "end": 3.0 }],
  "makespan": 5.0,
  "fitness": 5.0,
  "generations": 200,
  "fitness_history": [10.0, 8.0, 5.0]
}

Crossover types: "POX" | "LOX" | "JOX". Mutation types: "Swap" | "Insert" | "Invert".

GA config constraints:

Parameter Constraint Default
population_size >= 2 100
max_generations >= 1 200
mutation_rate 0.0 -- 1.0 0.1
tardiness_weight 0.0 -- 1.0 0.5
seed optional u64 random

Error handling: Invalid parameters return a JS error string (not a thrown exception). Check the return value:

try {
  const result = solve_jobshop({ jobs: [...], ga_config: { population_size: 0 } });
} catch (e) {
  console.error("Scheduling error:", e); // "ga_config.population_size must be >= 2, got 0"
}

Related