1 / 27

Planning and Scheduling Two Worlds in One System

Planning and Scheduling Two Worlds in One System. Roman Barták Charles University in Prague (Visopt B.V.) bartak@ktiml.mff.cuni.cz http://ktiml.mff.cuni.cz/~bartak. Talk Outline. Bridging the terminology gap Planning vs. scheduling Bridging the technology gap Example problems

Download Presentation

Planning and Scheduling Two Worlds in One System

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Planning and SchedulingTwo Worlds in One System Roman BartákCharles University in Prague(Visopt B.V.) bartak@ktiml.mff.cuni.czhttp://ktiml.mff.cuni.cz/~bartak

  2. Talk Outline • Bridging the terminology gap • Planning vs. scheduling • Bridging the technology gap • Example problems • Visopt ShopFloor System • From modeling to solving • Integrating planning into scheduling PLANET Information Day, May 26, 2003

  3. Terminology gapDo we understand each other? PLANET Information Day, May 26, 2003

  4. Planning and scheduling in practice • marketing planning • what is the company strategy? • long-horizon planning • what and when should be produced? • production planning • how to produce given items? • operation scheduling • what is the exact sequencing of operations? PLANET Information Day, May 26, 2003

  5. Planning task “The planning task is to find out a sequence of actions that will transfer the initial state of the world into a state where the desired goal is satisfied“ PLANET Information Day, May 26, 2003

  6. Planning task at example Final world C B A D C B A D Initial world C • Plan • Pick(C) • Release(C,table) • Pick(B) • Release(B,D) • Pick(C) • Release(C,B) B A D PLANET Information Day, May 26, 2003

  7. Planning task at glance • Input: • an initial state of the world • a description of actions changing the world • a desired state of the world • Output: • a structure of actions • Features: • Activities are not know in advance • No time and resources PLANET Information Day, May 26, 2003

  8. Scheduling task “The scheduling task is to allocate known activities to available resources and time respecting capacity, precedence (and other) constraints“ PLANET Information Day, May 26, 2003

  9. Scheduling task at example PLANET Information Day, May 26, 2003

  10. Scheduling task at glance • Input: • a set of activities with precedences • a set of available resources • Output: • allocation of activities to resources • Features: • Activities are known in advance • Time and resources are restricted PLANET Information Day, May 26, 2003

  11. Technology gapCan we integrate? PLANET Information Day, May 26, 2003

  12. Traditional „integration“ activities failure • Do planning first i.e. decide about necessary activities • Do scheduling next i.e. allocate activities to time and resources Actually, planning and scheduling is done separately there. PLANET Information Day, May 26, 2003

  13. Do we need to integrate more? Sometimes more tighten integration of planning and scheduling is necessary. When? • If there are too frequent backtracks from scheduling to planning. Improving the planner may help. • If existence of the activity depends on allocation of other activities. Foregoing planning of activities cannot be done there. PLANET Information Day, May 26, 2003

  14. Process-dependent activities heat re-heat process long duration by-product A B transition final product re-heating re-cycling heat process process setup PLANET Information Day, May 26, 2003

  15. Visopt ShopFloorIntegration in practice www.visopt.com PLANET Information Day, May 26, 2003

  16. Visopt ShopFloor at glance factory description customer demands(orders) CP plan/schedule • a generic scheduling engine beyond traditional scheduling PLANET Information Day, May 26, 2003

  17. Transition scheme produce A (3-4) produce B (1-2) produce C (2-4) load clean cool heat unload • scheduling resources with complex behaviour a state transition scheme with min/max batches A A A B C C C C A A A C C B A A A • batch counters • force given state (cleaning)after a given number of batches clean load heat unload load heat unload cool clean PLANET Information Day, May 26, 2003

  18. Dependencies supplier consumer dependency delay Alternative recipes Recycling N-to-N relations batches consume and produce items moved quantity items are moved between resources supplier-consumer dependencies • describe relations between batches (resources) PLANET Information Day, May 26, 2003

  19. Demands Production in the factory must be evoked somehow! This is realised via external demands. In Visopt ShopFloor the demands are described as orders of items: • item (product) • quantity • delivery time • alternative product? • variable quantity? • delivery delay? PLANET Information Day, May 26, 2003

  20. Optimisation Always ask why doesthe user need a given objective! What do the users need to optimise? minimise makespan minimise number of set-ups minimise lateness minimise earliness maximise resource utilisation ... more satisfied demands expensive set-ups penalty for delays storing cost fix expenses So what is the common objective? M O N E Y • In Visopt ShopFloor we minimise cost (= maximise profit): • income for delivered items (negative cost) • penalty for late deliveries • costs of production / moving / storing PLANET Information Day, May 26, 2003

  21. The task Input • factory description • resources • dependencies • initial situation • demands Output • planhow to satisfy the demands? • schedulehow to realise the plan? PLANET Information Day, May 26, 2003

  22. Slot representation time shift K-1 K+1 K end start end start duration Resource-centric model • a resource schedule = a sequence of batches • realised via a sequence of slots (slot = space for batch) • slots can slide in time • slots cannot swap the position • Slot variables • describe parameters of batch to be filled in the slot state = a batch type that can be filled into the slot serial = a position in the state sequence state +serial times = start, end, duration connected via constraints PLANET Information Day, May 26, 2003

  23. Dependencies K K-1 K-1 When the batch is known (located to a slot),ask for missing suppliers/consumers! Dependency generator(realisation of a simple planner)  Scheduling = deciding about batches in slots Using information about dependencies in the batch PLANET Information Day, May 26, 2003

  24. (Pseudo) Demo PLANET Information Day, May 26, 2003

  25. Conclusions PLANET Information Day, May 26, 2003

  26. A message of the talk • Planning and scheduling are different areas with different solving technologies. • But they can be integrated into a single solver! PLANET Information Day, May 26, 2003

  27. Planning and SchedulingTwo Worlds in One System Roman BartákCharles University in Prague(Visopt B.V.) bartak@ktiml.mff.cuni.czhttp://ktiml.mff.cuni.cz/~bartak

More Related