Search

To support the production rate, some transfer line stations with long cycle times must be duplicated. A job is processed in only one of the duplicate stations. If laid out in parallel, this duplication essentially doubles the capacity, but, for practical reasons, these duplicate stations are often laid out in a series. The serial layout has a capacity less than the parallel's, and also raises some operational questions. This actual problem of operating duplicate stations was enunciated and analysed in series. An existing closed-loop transfer-line was simulated with duplicate stations in series using the actual failure and repair characteristics of the automated stations, and compared a new heuristic with three simple policies. [ABSTRACT FROM AUTHOR]

One goal of the Toyota production system is a level usage of parts in assembly. To this end Toyota developed a 'goal chasing' algorithm for sequencing vehicles in its assembly plants. Since the goal chasing method I was computationally too burdensome for most real-world applications, Toyota developed a faster version called the 'goal chasing method 2.' While this modified version performs well in automotive assembly for which it was designed, it cannot sequence products with non-zero/one part requirements. For example printed circuit boards have non-zero/one product-part usage matrices. These and other electronic products can have any number of a particular chip, resistor, capacitor, or other electronic component. Hence the product-part usage matrix for these electronic products is not zero/one, and the goal chasing method 2 cannot be used. We present another modification of the goal chasing algorithm that is faster than Toyota's goal chasing method 2, and can sequence products with non-zero/one product-part usage matrices. [ABSTRACT FROM AUTHOR]

We extend an intuitive level part usage measure to the multi-level case, and provide a practical yet powerful algorithm The measure assigns an ideal due-date to each unit of each part type of each level, and sums the squared deviation between these ideal due-dates and the actual completion times. The algorithm schedules the final products in earliest ideal due-date order. This algorithm entirely ignores the lower levels and is consequently extremely fast. Surprisingly, in addition to being many times faster than other algorithms, this quick and dirty algorithm performs very well with respect to both the new ideal due-date measure, and the multi-level goal-chasing measure. [ABSTRACT FROM AUTHOR]

This paper considers a production environment in which multiple products produced in a batch process share several facilities, and the products must be grouped (or, equivalently, assigned to facilities) so as to minimize associated inventory carrying costs (an NP-hard problem). The main contributions of the paper are (1) to derive a polynomial-time procedure that computes optimal solutions to an important special case of the problem, (2) to extend the procedure as an heuristic to the more general case, and (3) to demonstrate computationally that the heuristic is remarkably effective in the general case. [ABSTRACT FROM AUTHOR]