| Budget Amount *help |
¥3,300,000 (Direct Cost: ¥3,300,000)
Fiscal Year 2001: ¥800,000 (Direct Cost: ¥800,000)
Fiscal Year 2000: ¥1,100,000 (Direct Cost: ¥1,100,000)
Fiscal Year 1999: ¥1,400,000 (Direct Cost: ¥1,400,000)
|
|---|
| Research Abstract |
Two critical, yet contradictory, concepts of modern production planning are grouping and smoothing. Batching of the same type of products to form batches or lots are essential for efficient shop operation when setup losses due to product changes are non-trivial. Smoothing has become viewed as an another key concept especially in the so-called JIT environment. These concepts have been studied individually but not together.
This study focused on a production line a part of which calls for batching (such as painting), whereas the other calls for smoothing (such as assenbly). The system considered is a two-stage hybrid flowshop in which the first stage consists of two parallel machines and calls for batching of products having the same attribute, say, color. Production smoothing, on the other hand, is a major concern of the second stage. A production sequence which "compromises" both batching and smoothing considerations are sought.
The study, after describing the problem scenario, presented a general idea to achieve smoothing and identified, under the assumpticn of "no-inventory synchronized production", properties of the production sequences to meet a smoothing "constraint". A mathematical program is fourmulated to obtain a final sequence to minimize the number of the first-stage setups due to changes of the product attribute (say, color). The study also discussed a method to make appropriate "color/product-type assignment" so that the number of color changes can be reduced. The problem is also formulated as an integer program. The proposed approach was evaluated computationally with real data.
The study also presented 1) a Lagrangian-based heuristic for the standard job shop scheduling problem where operation precedence constraints are relaxed to yield subproblems for each machine, and 2) a colum-gen aration heuristic for a lot scheduling problem of single-stage parallel identical machines.
|
