A scheduling system to achieve just-in-time production and minimum changeover counts in a polymerization plant

The operation of a polymerization plant is usually scheduled by being divided into blocks: a production block, a store block, and so on, with the operations of each block usually scheduled by experts. The production block is scheduled to minimize production costs. However, it is difficult to manually schedule the store block, because of the following features. (1) The products must be treated as continuums. (2) The products are concurrently handled by two or more machines for long periods. (3) The capacities of product queues have an upper bound. (4) Operations are not necessarily started when they can be. (5) The store block constitutes a very large system. This paper proposes a scheduling system to concurrently achieve "Just-In-Time" (JIT) production in the store block and to minimize changeover counts in the production block. The proposed system calculates the latest finish times for receiving using a pull type algorithm to achieve JIT production. The latest finish times and all combinations of sequences of orders in the production block are examined by a push type scheduling algorithm under restriction conditions of features (2) and (3). To schedule a very large system, an algorithm based on the branch and bound technique is employed. The effectiveness of the proposed system is assessed via numerical studies.