Scheduling (production processes) Study Guide

📖 Core Concepts Scheduling – arranging, controlling, and optimizing work and workloads in manufacturing. Purpose – allocate plant/machinery, plan labor and materials, and meet customer due dates. Goals – keep due dates, minimize production time, reduce costs (optimize when to make products, who works, and which equipment is used). Stochastic Scheduling – scheduling when any of processing time, due dates, weights, or machine breakdowns are random. Forward Scheduling – start from resource‑availability date and compute the shipping/due date. Backward Scheduling – start from the required‑by date and compute the latest feasible start date (and any capacity changes). Resource‑Task Network – a diagram showing tasks and the resources needed for each, used for analysis. 📌 Must Remember Forward → available → due date; Backward → due date → start date. Production scheduling lowers inventory, smooths labor load, and reduces change‑over time. Heuristic algorithms (e.g., Modified Due Date, Shifting Bottleneck) are used when exact optimization is computationally heavy. Stochastic algorithms such as Economic Lot Scheduling Problem (ELSP) and Economic Production Quantity (EPQ) handle randomness in demand or processing. Batch production scheduling = scheduling of grouped (batch) processes; closely related to finite‑capacity scheduling. 🔄 Key Processes Identify Inputs – plant facilities, labor, materials, tooling, energy, clean environment. Map Output Flow – output of one work area becomes input to the next (e.g., cut → bend). Choose Scheduling Method If you know when resources start → use Forward Scheduling. If you have a firm due date → use Backward Scheduling. Apply Algorithm Large task set → select a heuristic (Modified Due Date → prioritize by adjusted due dates; Shifting Bottleneck → resolve bottlenecks iteratively). Random elements present → use a stochastic algorithm (ELSP, EPQ). Generate Visual Tool – create a Gantt chart for timeline view; optionally use Kanban signals for pull‑based production. 🔍 Key Comparisons Forward vs. Backward Scheduling Start point: resource availability ↔ required‑by date. Outcome: predicted shipping date ↔ latest start date. Heuristic vs. Stochastic Algorithms Heuristic: rule‑of‑thumb, fast, good for large deterministic problems. Stochastic: incorporates randomness, yields optimal batch sizes under uncertainty. Batch Production Scheduling vs. Finite‑Capacity Scheduling Batch: groups of identical items, common in pharma/chem. Finite‑Capacity: any limited‑resource environment; techniques overlap. ⚠️ Common Misunderstandings “Scheduling = just making a timetable.” – It also optimizes resource use, inventory, and labor balance. Confusing forward scheduling with “earliest start.” – Forward scheduling starts at resource availability, not necessarily the earliest possible start. Assuming heuristics always give the best solution. – Heuristics are fast approximations; they may miss the true optimum. 🧠 Mental Models / Intuition “Chain reaction” model: Think of each work area as a domino; the output of one is the input to the next. Scheduling ensures each domino falls at the right time. “Time‑budget” view: Treat the due date as a budget you must spend backwards from; any spare time can be re‑allocated to earlier tasks (backward scheduling). 🚩 Exceptions & Edge Cases Random machine breakdowns → switch from deterministic heuristic to a stochastic algorithm. Very tight due dates → backward scheduling may reveal infeasibility; may need to add overtime or extra capacity. Batch size constraints (minimum/maximum batch) → standard EPQ formulas need adjustment; use ELSP for stochastic batch sizing. 📍 When to Use Which Forward Scheduling – when start dates are known (e.g., after a new shift begins) and you need to predict delivery. Backward Scheduling – when the customer due date is fixed and you must determine the latest feasible start. Heuristic (Modified Due Date, Shifting Bottleneck) – for large deterministic job sets where quick, good‑enough schedules are acceptable. Stochastic (ELSP, EPQ) – when processing times, demand, or breakdowns are random; you need optimal batch sizes under uncertainty. Kanban – when you prefer a pull system driven by actual demand rather than a fixed schedule. 👀 Patterns to Recognize Bottleneck appears repeatedly → consider Shifting Bottleneck heuristic. Due dates clustered tightly → Modified Due Date heuristic often improves lateness metrics. Large number of identical tasks → batch scheduling or EPQ likely applies. Resource constraints visible on Gantt chart (overlap) → indicates need for capacity smoothing or additional resources. 🗂️ Exam Traps Choosing forward vs. backward based solely on “earliest start” wording – read the prompt: available date ⇒ forward; required‑by date ⇒ backward. Assuming “heuristic = inaccurate” – many heuristics are proven to be within a small % of optimal; they are often the correct answer for large‑scale problems. Mixing up batch scheduling with single‑machine scheduling – batch concerns grouped production; single‑machine focuses on order of individual jobs on one resource. Kanban vs. Gantt – Kanban is a pull signal system, not a timeline chart; Gantt is a visual schedule. --- Use this guide for a rapid refresh before the exam – focus on definitions, when to apply each method, and the typical patterns that signal which tool to reach for.