1. BUCKET BRIGADES An Approach to Autonomous Work Allocation in Flow Production Systems (Reading No. 7 in poly-copy) Also INDEX snippet on course site

2. Conventional Approach (Order Picking) Products Order #xyz 2 blue jeans 3 red hats 1 green shirt Order/product flow Completed orders Customer order “Zone picking”

3. Conventional Approach (Order Picking) Order/product flow Products Order #abc 10 green shirts 4 yellow ties Completed orders Customer order “Zone picking”

4. Bucket Brigades Order/product flow Products Completed orders Customer order In bucket brigades, there are no zones. Individuals continue to add to an order until it is taken off them Order #xyz 2 blue jeans 3 red hats 1 green shirt Meet and handoff

5. Bucket Brigades Order/product flow Products Completed orders Customer order In bucket brigades, there are no zones. Individuals continue to add to an order until it is taken off them Order #xyz 2 blue jeans 3 red hats 1 green shirt

6. Bucket Brigades Order/product flow Products Completed orders Customer order In bucket brigades, there are no zones. Individuals continue to add to an order until it is taken off them Meet and handoff Order #abc 10 green shirts 4 yellow ties

7. Bucket Brigades Order/product flow Products Completed orders Customer order In bucket brigades, there are no zones. Individuals continue to add to an order until it is taken off them Order #abc 10 green shirts 4 yellow ties

8. Product flows down line in overall progression Usually more stations/ workplaces than workers, who move to where the work is to be performed Each worker takes product towards completion along the line, and when last worker finishes product: – Despatches it and walks back to take-over work from immediate upstream worker, – who in turn goes back to preceding worker, and so on Works as a continuous looping relay with both ways movement for baton pass How Do Bucket Brigades Work? Middle of page 2 in reading

9. How Bucket Brigades Work: A Simple Animation

10. Work mostly comprises variations on a few/ simple skills, and is spread along the line Easy/ fast movement between stations Easy/ fast take-over of work between workers Stations are inexpensive w.r.t. labour costs Preferred sequence Slowest-Fastest No passing, alteration of worker sequence, or accumulation of buffer WIP permitted Assumptions/ When Appropriate Middle of page 7 (& 2) in reading

11. Inspirations from Natural Systems “Models of organization that are based on living systems are naturally organic and adaptive.” Most biological systems do not forecast or schedule: They respond — quickly, robustly, and adaptively The Conceptual Difference: Rather than trying to control the system, Design the system so that it controls and adapts itself.

12. Nectar Collection in Honey Bees ForagersReceivers May split load up to 6 times!

13. If waiting time is long: recruit more receivers If waiting time is short: recruit more foragers If waiting time is intermediate : do nothing No need to know details about: – supply and demand – rate of arrival – number of other bees queueing But can track a fluctuating and unpredictable environment Work Organization Simple individual level rule (whose?):

14. Food Foraging & Transport in Ants NestFood

15. Bucket Brigades Order/product flow Products Completed orders Customer order NestFood

16. Several Toyota subsidiaries/ suppliers (>100) since 1970’s; outside Japan since 1989 Toyota Sewn products System (TSS) developed by Aisin Seiki for seat-covers – Each worker takes over sewing of predecessor’s item – Usually based on standard workers with possibly varying processing requirements at different stations Champion Products for sewing Couch Leatherware for sewing Manufacturing Applications: Apparel/ Upholstery

17. TSS Snapshot

18. U-shape to reduce walking/ improve communication Work (sewing) done standing up at waist-chest high stations Number of workers generally preferable < 12: – Ex.: 2 workers making T-shirts (2.5’ work content) – Ex.: 16 workers making Leather covers (190’ work content) Reset times (returning to all-forward movement) as low as 3-10 secs. in a 3-worker/ 13 station line with 7’ work content Sometimes a little work squeezed in during walk-back too Layout & Configuration Aspects

19. Mitsubishi Consumer Electronics America for TV assembly, Cell-phone packing United Technologies Automotive for automotive harnesses Tug Manufacturing for assembly of airport tractors Subway for sandwich assembly (at 13,000 franchisees) Manufacturing Applications: Simple Assembly Jobs

20. Single most fertile area, involving much walking/ picking but less dexterity For high-volume Distribution Centres (DC’s) – Orders arrive at beginning of aisle of multiple rows of flow racks – Pickers pick items from a segment of rack along conveyor, adding until next person takes over Order variations, load fluctuations etc.: difficult to plan/ balance statically Warehousing Applications: Order Picking in Distribution

21. Warehouse Pick Line Snapshot

22. McGraw Hill DC’s Time-Warner Trader Publishing/ Little-Brown Bantam Double Day-Dell Distribution Reader’s Digest (8% rate increase, 35% error reduction) The Gap DC’s serving Old Navy, The Gap & Banana Republic Stores (throughput up 25%) Revco Drug Stores (now CVS) in national DC’s Warehousing Application Examples

23. Less need for detailed time study, pre-planning and accurate balancing Simple, identical protocol which still enjoys repetitiveness benefits Only limited multi-skilling needed (neighbours) Pull system: limited WIP but can cope with changes Manual handling often sufficient Consistent with other contemporary small-group based Operations approaches Potential Benefits

25. From polycopy reading no. 7 (Section 2.1 and Appendix B) Each worker (i = 1…n) is characterized by a work rate, based on the amount of time required by that worker to complete one full unit, and a (relative) work velocity v i w.r.t. a “standard” worker Workers are arranged slowest to fastest To calculate aggregate production rates and steady-state “operating zones” Quantitative Analysis of Bucket Brigade Lines

27. Dropping off work in buffer if station busy – Workers keep circling pre-bottleneck => fewer available to clear bottleneck and post-bottleneck – When worker leaves bottleneck station, fresh worker may not be available => bottleneck idle – Need to plan number of workers more rigorously, defeats self-planning objective – Can only protect by buffering worker/station too Possible Improvements?

28. Workers circling around in order to reduce hand- off delay – If workers speeds are not same, all workers will accumulate behind slowest worker (no-passing case) – If passing allowed, faster workers protected but slower workers may need to wait for stations to get free => workers not used fully – Also less learning benefits, since performing entire set of tasks rather than in own zone and its vicinity Possible Improvements?