A SCHEDULING APPROACH TO BATCH PRODUCTION ENVISAGING THE REDUCTION OF JOB THROUGHPUT TIMES

Ana Almeida, Carlos Ramos
GECAD - Knowledge Engineering and Decision Support Group & Computer Science Dept. Institute of Engineering – Polytechnic of Port, Rua Dr. António Bernardino de Almeida, 431 4200-072 Porto, Portugal Email: ana@dei.isep.ipp.pt, csr@dei.isep.ipp.pt

Abstract This paper is concerned with detailed scheduling of job-shop like manufacturing systems. It addresses the scheduling of jobs, either simple, requiring the manufacture of a batch of parts, i.e. simple products, or complex, comprehending the parts fabrication and their multistage assembly into a batch of products. The Product-oriented Scheduling approach was adopted, assuming that full scheduling of a simple or complex job, based on the job routing network of operations, from the first operation to the last, is performed before another job is considered for scheduling, having in consideration existing manufacturing processors and their availability. We followed this approach because we aimed at compressing job flow time to a minimum as a strategy to meeting job due dates. To further enhance this objective the idea behind Simultaneous Manufacturing was implemented. In particular, the widespread use of batch overlapping was implemented with Job Scheduling Patterns, which proved particularly effective in reducing job throughput time, maintaining operating simplicity and requiring reduced coordination. Key words Manufacturing Systems, Batch Production, Simultaneous Manufacturing and Product-oriented Scheduling.

1 INTRODUCTION To know what work to do at each moment in time and using what resources is a central need to every company. The answer to this lies in good scheduling, which strongly contributes for the success of a company manufacturing operations. Scheduling means solving two problems, namely that of allocation of resources to each task and that of determining the moment when each task will be performed, (Baker 1974). Real life scheduling problems tend to be so complex that, usually, optimal solutions cannot be found. It is widely accepted that manufacturing scheduling problems are generally difficult or hard to solve to optimality (Baker, 1974), (French, 1982) and (Blazewicks, 2001). This has not been an argument to abandon efforts for finding methods that can achieve optimum or near optimum solutions, since much work appeared, during the last decades, on this matter. However, for most real world problems, found in practice, this can be a very difficult and, many times, an impossible job. This is so because the number of variables tends to increase in relation to theoretical formulated problems, which are already hard and, additionally, because the environment is predominantly dynamic and non-deterministic (French, 1982), which makes things even more difficult. This means that, in a given moment, an established schedule, even an optimum one, for some particular scheduling instance, soon becomes invalid and may have to be scraped or

1

2 PROBLEM DEFINITION The problem under consideration can be included in the job-shop class with renewable resources. called simple products. for this type of industries that the scheduling approach and methods presented in this paper are most appropriate. every time scheduling is to be done. in any manufacturing processors setting. have also adopted this name. the methods are of reduced practical value. The scheduling approach combining all or some of these factors for reducing job throughput time has been named as Simultaneous Manufacturing (Silva. problems are of different nature. 1997). frequently. 1995). We try to address practical problems. job-shop like environments. the manufacturing operations and their precedence relationship. the manufacture of an ordered quantity of identical chairs. each one made of several different parts. The most common academic approach to scheduling is to consider static and basic problems (Portmann. but not only. again. having a set of operations carried out in a given sequence in a batch of products. 1996). However. for example. which most probably would not be possible. they are more general and complex than the already very complex job-shop scheduling problem (JSSP). designed for getting good or acceptable solutions. i. The products may also be complex products.adapted. Because it assumes the existence of alternative resources to process the task we can say that is a 2 . The manufacture of a certain amount of a specific component of a chair is not a job. their agendas. The manufacture of these complex products is typical of industries of type A and T as defined by Chase and Aquilano (1998). involving frequently simple mechanisms. we focus the quality of scheduling in shortening job throughput times and meeting job due dates. Therefore. A typical job is. To further enhance this multi-objective we adopt a strategy to scheduling known as horizontal scheduling (Vollmann. Because a job is generally associated with a product. in the short available time for taking decisions.e. methods for such uncertain environments and complex problems usually tend to be of heuristic nature. we preferred concentrating on developing a strategy and methods to efficiently schedule jobs. We strongly explore the possibility of different batch overlapping and batch splitting schemes for finding good schedules. Under horizontal scheduling. a job is schedule first in all the required processors before another job is considered for scheduling. we decided to name this scheduling strategy as Product-oriented Scheduling (POS). for which methods are developed. tested and compared. The reason is that this has a positive contribution to both profits and customer service. It is mainly. is typical of reality but untypical in theoretical scheduling problems addressed in the literature. Our approach considers a job to be a manufacturing order for a number of identical products. because in practice. i. The most generic job requires several operations on identical and different parts followed by assembly operations. priority to execute a job already started is given and consequently flow time further reduced.e. These may be simple. when alternative manufacturing processors exist batch splitting may become a specially attractive strategy for shortening flow times. For this we use both simple priority rules and some analytical procedures. made of fabricated parts followed by assembly. having a set of operations to be carried out in a given sequence in a batch of identical parts. rarely optimum ones. instead of concentrating on developing optimum methods. its batch size. While the former is easy to implement and can be of great impact in reducing job flow times. such as Hastings and Yeh (1990). However. This is the reason why. Therefore. Actually. i. A job as defined in this work. in similar manufacturing settings. Processing times include auxiliary time elements such as set up time and time for parts handling at a machine.e. processing times of operations and the manufacturing processors to be used. which take into account actual time availability and unavailability of manufacturing processors. Other authors. A job is specified by the quantity of products necessary.

SM philosophy does not guarantee the most economical way of manufacturing and order.e. We use concurrent scheduling in shop floor without constraining schedules to lead times. the product order. POS is used by ERP software for capacity planning based on lead times. having in consideration that jobs are manufacturing orders of products. it is more general than this because any task may be processed on more than one machine at the same time. As we can envisage SM calls for intensive batch overlapping and for batch splitting under alternative processors. The main purpose is to reduce job throughput time as a strategy to meet due dates.1 Product-oriented Scheduling Product-oriented Scheduling (POS) is identical to horizontal scheduling (Vollman et al. This assembly must be performed immediately. typically addresses in JSSP literature. in such a way that all the operations are scheduled in all the required re-sources or processors before the next job is considered for scheduling. To achieve this. In fact. and therefore the full job throughput time is minimum too.. This scheduling problem scenario includes a much general case than the basic JSSP. which permits verifying if job due date is likely to be accomplished. its throughput time may be very small indeed. POS implements the concept of Simultaneous Manufacturing in batch production of complex product manufacturing orders. Using this approach it is possible to have a good perception of the state of each job during the scheduling time horizon and easily establish the scheduled job completion date. under simultaneous widespread use alternative processors by the same product order. In this way. to be fabricated and assembled.flexible job-shop problem (Brucker and Schlie. each set of parts belonging to each product of the order must flow in a coordinated way through the system. resulting from fabrication of parts and their assembly. We call the scheduling problems of the type defined as Extended Job-Shop Scheduling Problems (EJSSP). The scheduling process is focused on the job. belong to the same task. Besides that. i.2 Simultaneous Manufacturing The Simultaneous Manufacturing (SM) philosophy aims at the complete manufacturing of each single product of a product order in the minimum possible time. 3. it is considered a variable number of operations per task. 2003). These are variable in time and depend on job and processors availability. Additionally. and arrive simultaneously to where they are needed for assembly. 3 ADOPTED APPROACH TO SCHEDULING 3.e. i.. mechanisms were developed for implementing SM in a usercontrolled manner. so it is not restricted to the pure job-shop problem but it refers to the general job-shop problem (Conway et al.1990). 2000). 1996). The intention is to take the minimum time to manufacture the whole job.. This happens not only because several parts. (Mastrolili and Gambardela. the work-in-process is likely to be low. in a way such that they are processed before any other parts. through the widespread use of batch overlapping. and is more realistic and in line with what happens in practice. In our scheduling approach. according to available processors. In this work this is achieved. but set-up costs tend to be high (Almeida et al. the throughput time for the complete manufactured and assembly of each product of the product order is minimum. 1967). Usually. Additionally. 4 BATCH PRODUCTION 3 . but also because batch overlapping and batch splitting strategies may be adopted.

i.e. This means that batch processing and batch transfer should be linked in a manner that each transfer batch can be transferred to its successor operation. are not critical. An apparent disadvantage of this strategy is the overall increase of batch set-up times.2 Batch Overlapping Batch overlapping means transferring work from a machine. normally smaller than the total job batch size.e. this can become a too great penalty to the full duration of the job processing. transfer batches are clearly defined. contributing for improved product delivery. When a transfer batch is equal to the total job batch size batch overlapping does not take place. the transfer of work between processors should be continuous. This may not be important. are not bottlenecks. which allow implementing POS and SM in our work. In our research.1 Batch Splitting Batch splitting means partition of a batch into several smaller batches to be processed independently on machines. This guarantees low job throughput time. a substantial reduction on average batch flow time may be achieved. we developed a mechanism for avoiding the existence of these inactive or idle times based on flexible batch transfer sizing which is the same to say that different transfer batch sizes must be determined between succeeding adjacent job operations. based flexible batch 4 . in practice. In this way. batch overlapping and batch splitting. Therefore. which is processing an operation of the job. In cases where the batch size is large. This is very common. Batch overlapping requires combination with batch streaming. Batch overlapping does not necessarily change the processing batch size at a processing stage. 4. with different amounts of overlapping. mainly regarding job throughput time and accomplishment of job due dates can become highly poor. a job is considered as a set of identical parts that are always processed as a whole. This coupling calls for a more close control of production. A JSP is a virtual schedule of a given job. 4. and other times under a well defined overlapping procedure. splitting a batch into multiple transfer batches for overlapping. transferred between two successively required machines. This operating weakness can be highly reduced through Horizontal Scheduling. if processors. when trying to fully implement Simultaneous Manufacturing we should seek maximum overlapping. couples machines at different stages. A processing batch size is the amount of units of a job processed in a machine continuously before it takes another job. the full batch must be processed in a processing stage before it can be transferred to another to carry out further processing. In the extreme. This may be particularly so when alternative machine or processors exist for simultaneous manufacture of split batches. We explore these strategies in our approach to solve the EJSSP. Additionally. sometimes done randomly. intermittent idling may exist at machines whenever a predecessor operation requires a longer processing time. if products arrive at machines in identical transfer batches. In this case. for processing the next operation. However. immediately upon completion. several transfer batches can be concurrently processed at different job operations. i. before the entire batch has been finished on the previous machine. These are batches. neither be pernicious to job flow times. to another machine. eventually requiring an increased effort in rescheduling under disturbances. and is likely to reduce work-in-process inventory.e. which are set-up. So the performance of the manufacturing system. i. is what we call Job Scheduling Patterns (JSP). 5 JOB SCHEDULING PATTERNS One of the main mechanisms.In traditional batch production. which means transfer batches of size one.

transfer sizing. 5 . simple or complex. in an unspecified time horizon. This specifies all operations and their precedence interrelationships of each single product. i. and last level operations are those. It may be expressed as graph as the one shown in figure 1. In our work both of them are considered. 2002). 5. where b = succ(a) a ff b represents a precedence relationship. we describe the mechanism used to generate JSP. In this section. of a job order. A job can have several JSP. for a given job. The other operations are called internal operations. with all processors considered available.1 Notation used prec( ) precedence operator succ( ) succession operator a and b represent two different subsequent /succeeding operations of the same job k to be processed at different manufacturing processors. we consider the operations without precedents as operations of level one. Two types of graphs can represent an operation plan namely an outree and intree graph French (1982). meaning a directly precedes b rk job release date tpa operation a processing time for one product unit tpb operation b processing time for one product unit ltp(k) job k processing batch lt(a ff b) transfer batch between a and b Δts(a ff b) displacement or overlapping time between the starting time instants of two Δtf(a ff b) displacement or overlapping time between the finishing time instants of two operations a and b operations a and b tstart(a) instant of processing start for operation a tfin(b) instant of processing end for operation a tstart(b) instant of processing end for operation b tfin(b) instant of processing start for operation b 5.e. A JSP is one of several possible alternative schedules. which do not have succeeding operations. batch overlapping and/or batch splitting (Almeida. op1 Op2 Op4 op3 Figure 1 Intree precedence graph representation of a Job operation plan An important concept is the level of the operation. under an empty manufacturing system.2 Operation plans A requirement for establishing a JSP is the job operation plan. which can be generated for better exploring the utilization of manufacturing processors or machines.

…. sum up the processing times of all operations: tpCl = where: opi =1 ∑ tp n opi . on the job operation plan adopted and on machines to be used. (1) opi ∈ Cl Cl with l=1.3.5. c represents each graph path prec(op1) = {} ∧ succ(opn) = {} 6 .Execute Step 4 store the obtained values at the list of processed operations For the non-critical paths From the last level operation until the operation of level one If opi ∉ to the highest duration graph path make a = opi and b = succ(opi) compute tfin(a) e tstart(a) – Execute Step 5 compute the displacement Δts(a ff b) .3 Computation of JSP The computation procedure determines the starting and finishing time instants for all the operations of a job on each processor or machine depending on transfer batch size. Ordering and storing of operations for each graph path For each graph path. This establishes a JSP. 2.Execute Step 6 store the obtained values at the list of processed operations 5. The computation procedure is firstly applied to critical paths and then to other paths.2 Computing Procedure Steps Step 1 – Computation of the processing time for all the graph paths.Execute Step 3 compute tstart(b) e tfin(b) .1 Overall computing procedure Compute the duration of each graph path and store the sequence of operations for each one – Execute Step1 For the critical paths From the operation of level one until the last level operation If prec(opi) = {} and opi is the first operation (level one) Then compute tstart(opi) and tfin(opi) – Execute Step 2 store the obtained values at the list of processed operations Else make a = opi and b = succ(opi) compute the displacement Δts(a ff b) . A critical path has the longest duration of operations from the start node of the graph to the last. To begin we need to distinguish between critical paths and non-critical paths in the graph of the operation plan.3. 5.

namely through the following expression: Δts(a ff b)= lt(a ff b) * tpa (4) 7 . i. In order to simplify the computation procedure we consider that n/lt(a ff b) ∈ . where: rk its the job release date tfin(a) = tstart(a) + n * tpa (2) (3) Step 3 – Computation of the displacement or overlapping time between two succeeding adjacent operations a and b (Δts(a ff b)) from critical paths The batch displacement between two succeeding adjacent operations a and b. is obtained.lt(a ff b) )* tpa This means that the total processing time for all the units of the batch for the operation b is long enough to ensure that no interruption occurs during batch processing at operation b. The transfer batch can. Figure 2 a) illustrates the relationships between important variables in such a way that one can easily understand how the displacement for this case. is the lowest time interval between the starting of the two operations that ensures enough feeding to the machine which processes operation b for avoiding processing interruption.e. This displacement allows defining the transfer batch lt(a ff b). So the processing of operation b can start as soon as the transfer batch arrives at the respective manufacturing processor. so the computation mechanisms are different. the size of the transfer batch depends on factors such as the minimum and maximum amount work that can be transferred between the manufacturing processors and the buffer size of the receiving manufacturing processor or machine. 2. after the first transfer batch has been processed in operation a. c represents each graph path Step 2 – Computation of the starting and finishing instants for the first operation in critical paths tstart(a) = rk . therefore.tpopi it is the processing time of operation opi tpC it is the jobs total processing time for the graph path under consideration l The following set is obtained: TpC = {tpC } l l where: Cl with l=1. …. Considering a job batch size of n units. Two cases can occur depending on operation processing times. Case a) If tpa ≤ tpb then tpb * n ≥ lt(a ff b) * tpb + (n . of the same job. be unitary or have a size not exceeding the size of the job batch size. intermittent idling.

illustrates how such displacement is obtained in this case. the displacement must be long enough to ensure processing of operation b without interruption. mp2 lt (a ff b) *tpb δ n*(tpa −tpb) D δ Δts(a ff b) a. Thus. is: Δts(a ff b)= n*(tpa .tpb )+ lt(a ff b) * tpb (5) Step 4 – Computation of the starting and finishing instants for the other operations of critical paths (6) tstart(b) =tstart(b) + Δts(a ff b) where a is the operation with the greater starting instant (tstart).≥ lt (a ff b) *tpb a. mp1 lt (a ff b) *tpa lt (a ff b) *tpa Δts(a ff b) a. in the universe of all the preceding operations of b.lt(a ff b) )* tpa so: tpb * n < lt(a ff b) * tpb + (n . as illustrated in figure 3. from figure 2 b) we can see that δ = lt(a ff b) * tpb In fact. then it would occur a batch processing interruption due to starving.lt(a ff b) * tpa (8) 8 . the displacement for this case. mp1 a. Therefore. intermittent idle time would take place. after operation a has been processed.lt(a ff b) )* tpa This means that the total processing time for all the units of the batch for the operation b it is not long enough to avoid interruption during batch processing.tpb )+ δ+ n* tpb it comes δ = lt(a ff b) * tpb Therefore. mp2 lt (a ff b) *tpb Figure 2 Displacement representation for Step 3 a) when tpa ≤ tpb b) when tpa > tpb a) b) Case b) If tpa > tpb then tpb * (n . tfin(b) = tstart(b) + n * tpb (7) Step 5 – Computation of the starting and finishing instants for the all the operations of the other paths Case a) If tpa ≤ tpb tstart(a) = tstart(b) .e. i. Figure 2 b).lt(a ff b)) < (n . So if the processing of operation b starts as soon as the transfer batch arrives at the respective manufacturing processor. being: D = n*tpa+ lt(a ff b) * tpb and D = n*(tpa .

tfin). (op1. the result obtained by the job scheduling pattern generator mechanism.. . td. being the release date.. tfin)). and between op2 and op4 is shown.tstart(b) Case b) If tpa > tpb Δts(a ff b)= tstart(b) . rk. belonging to a non-critical path. Case a) If tpa ≤ tpb Δts(a ff b)= tstart(a) . tfin))} 5.n * tpa (9) (10) (11) When operation a also belong to the graph path with the highest duration its starting and finishing time instants are already know.lt(a ff b) * tpb tstart(a) = tfin(a) .. (op1. In table 1 are represented the processing times.Case b) tfin(a) = tstart(a) + n * tpa If tpa > tpb tfin(a) = tfin(b) . the preceding and succeeding operations as well as the occupation times for each manufacturing processor. Time tp opkij 1 3 1 2 Occupation time (n*tp opkij) 8 24 8 16 Prec(opki) op1 op2 . tfin). (mpn. . is illustrated in figure 3 where the batch displacement between op1 and op2. (opn.. op3 Succ(opki) op2 op4 op4 - Considering the data of the example and a unitary transfer batch.4 Illustrative Example We consider a very simple example in order to illustrate the exposed procedure. the instant 0. the respective manufacturing processor for each operation. of 70 tu. mpj mp1 mp2 mp3 mp4 Proc.Supposing that we intend to manufacture a batch with 8 product units within a deadline. td = 70 and rk = 0. we have: n = 8. tstart. (opn. so in this case it is not necessary to compute those values again.. Table 1 Processing times and operation precedence Operation opki op1 op2 op3 op4 Manufact.. ..tstart(a) (12) (13) We adopt the following representation for a JSP: JSPk ={(mp1. tstart. Processor... 9 . The precedence graph is represented in figure 1. tstart. Step 6 – Computation of the displacement Δts(a ff b) between the starting time instants of two succeeding adjacent operations a e b. tstart.

The starting and finishing instants of the occupation periods of the processors became different although the amount of time involved is the same. m p 2 op 2 ) * tp op 1 lt ( op 2 f f op 4 ) * tp op 4 o p 3. 10. (mp4. m p 3 > lt ( op 2 o p 4.(op3. (mp2. (mp3. If.(op1. 2002). instead of a unitary transfer batch. 27))} We point out that operation op3 could start anytime between instants 0 and 10. we consider. 48))} It can be noticed that the job flow time has increased from 27 tu to 48 tu. (mp4. which implements the horizontal approach generates a scheduling solution. 6 SCHEDULING THROUGH JSP With basis on the JSP and processor agendas the Scheduling Plan Generator (SPG). a transfer batch with the same size of the job batch the resulting job scheduling pattern is different and shown figure 4. 8 op1.(op4. 25)). The SGP draws upon a scheduling algorithm developed in previous work by the authors (Almeida et al. 8)).. 8)).(op1. 11. (mp3. 32)).(op3. By starting at instant 10 a late start JSP is used. m p 1 op 2 ) * tp op 1 ≥ lt ( op 1 f f op 2 ) * tp op 2 lt ( op 1 f f o p 2. 32. 0.(op2. in the other extreme. 25)). 8. m p 4 1 f f op 4 ) * tp op 2 10 11 18 25 27 8 Figure 3 Job Scheduling Pattern with batch displacement representation JSPk ={(mp1.(op4.(op2.n − lt ( op 1 ff o p 1. mp4 32 48 Figure 4 Job Scheduling Pattern with transfer batch equal to processing batch JSPk ={(mp1. mp1 8 op2. 0. 18)). 1. mp2 24 op3. Such algorithm involves two scheduling phases: the forward influence phase and the backward 10 . If had started at instant 0 we were adopting the earliest start JSP. (mp2. mp3 32 32 op4. 24.

Figure 5 shows the agendas of four manufacturing processors required for processing the job referred above. 5 pf 1 15 30 45 60 70 14 pf 2 11 pf3 8 0 23 20 38 45 55 50 pf4 50 60 Figure 5 . op2 11 mp3. 2002). are represented in figure 6 for the forward influence phase. when it is not possible to schedule the whole job batch to meet its due date. Another important feature the SPG. In the scheduling mechanism. the obtained results for each phase of the algorithm. pin pointing those intervals where manufacturing jobs can be successful and efficiently scheduled. 5 mp1. It verifies the time horizon of the processors associated to each operation and looks for their availability in a relevant time interval. op4 Occupied time intervals Occupied time intervals by the task Available time intervals 15 30 45 60 70 45 55 20 38 50 61 23 26 47 50 60 Figure 6 .Results of application of the Scheduling Plans Generator Going back to the example. The results for the backward influence phase are shown in figure 7. 11 . considering a unitary transfer batch.influence phase. implemented by the SPG. Mechanisms for implementing these phases were re-ported and explained in (Almeida. related with the scheduling problem and suggests possible solutions if they exist. within the available time intervals of the manufacturing processors. being this the final result for the scheduling of the referred job. which means using the JSP shown in figure 4. This may be important for time phasing delivery and negotiation with customers. It represents the MP available time intervals for scheduling jobs. represented by the black time intervals.Agendas of Manufacturing Processors The SPG picks up information about the work and processors. op3 8 mp 4. One of the strong points of the SPG is its capability of. is its ability to suggest a split batch size that can be manufactured within the due date. the concept of a manufacturing processor (MP) agenda is important. op1 14 16 mp2.

It seems. Thus. By relaxing batch size. on the backward influence phase. When it is not possible to schedule a job within the processors agendas. smaller job batches whose sizes can fit within the avail-able time intervals of the manufacturing processors. op2 11 mp3. to envisage the implications of decisions taken on the whole performance of the system. such as those of industries of type A or T (Chase and Aquilano.Results of application of the Scheduling Plans Generator It can be observed that. within the available time intervals in the manufacturing processors agendas. 1998). based on a given JSP. trough batch splitting. as a consequence poor solutions are likely to be obtained. and applying a batch splitting process followed by JSP generation for the split batches. considering different transfer batches sizes. for example. the time interval [20. scheduling relies solely on the ability of schedulers on the shop floor. therefore. op1 15 28 30 45 60 70 14 16 mp2. we can break one job batch in two or more. Frequently. In this case the split batches of the order could be taken separately for scheduling by the same method and. alone and without tool aids for scheduling. So.38] on mp2 has been limited to [25. These strategies can include trying new JSP. op3 8 mp4. a slack of 5 tu in each MP. It simply it is not possible for a person. further thinking is necessary to choose among alter-native schedules. then some strategies can be applied to obtain a solution. The scheduling mechanisms in the SPG give suggestions about where to split a job batch.47]. under complex manufacturing environments (referring to complex products). However. 12 . possibly provide further alternatives to scheduling.47] on mp4 has been limited to [26. it might be possible to take advantage of other available time intervals of manufacturing resources. It exists at the beginning of each useful time interval. to be important that tools are provided. the time interval [50.38] and the time interval [25. op4 23 26 20 25 38 45 55 50 47 50 60 Figure 7 .28]. 7 CONCLUSIONS The main concern in scheduling in practice is to obtain good solutions to schedule jobs at hand. the time interval [15. most probably the obtained solutions are less than good. and. due to the lack of aids for this.5 mp1. If more than a single set of intervals exist for scheduling.30] on mp1 has been limited to [15. some additional criteria and heuristics may be put on top of the scheduling mechanisms used to arrive to improved solutions.61] on manufacturing processor mp3 has been eliminated. which can improve the task of schedulers. These slack times can be used to adjust forward or backward the schedule The only solution obtained stems from some im-posed restrictions on batching.

. Ana. 2003. L. Artiba and S. “Effective Neighborhood Functions for he Flexible Job Shop Problem”... Pesch. keeping work in process low. Springer. (1990). The approach uses the Job Scheduling Pattern concept. “Simultaneous Manufacturing in Batch Production”. Doctoral Thesis on Production Engeneering. N. 369-375. New York (2nd Edition). Ramos. "Toward Dynamic Scheduling of Manufacturing ". W. edited by A. French.. K. Chapman & Hall. (1974). K. This has a direct effect on profitability. Silvio Carmo (2002). ISATP2003.. Portugal. University of Minho. R. “Introduction to Sequencing and Scheduling”. “Theory of Scheduling”. and Yeh. which can be seen as a powerful tool to aid users to improve detailed shop floor scheduling in complex manufacturing environments. 3-20. Computing 45. C. nº47. Homewood. and. (1982). Conway.A. S. & Silva. Irwin. 3. Mastrolilli. REFERENCES Almeida. (2002). Portmann. reduce costs in manufacturing. One clear contribution on these lines is to manufacture job orders within the shortest time that is reasonably possible. Chase. The Planning and Scheduling of Production Systems. N. and implements a Product-oriented Scheduling approach in order to achieve the Simultaneous Manufacturing philosophy. P. AddisonWesley Publishing Company. H.Moreover. (2001). since fast turnover of short-term investment is achieved. Schlie R. E. Portugal. L. (2003). R. which is an important requisite for customer satisfaction. Ana. and Miller. on fast deliveries. R. Almeida. M. “Scheduling methodology: optimisation and compu-search approaches I”. Besançon. C. and Jacobs. Proceedings of the IEEE International Symposium on Assembly and Task Planning. F. John Willey & Sons. E. (1990). Schmidt. New York: Halsted. (1997).. A. Methodologies and applications. submitted to the University of Minho. Based on the described scheduling approach a Scheduling Plan Generator was developed. Aquilano. Vol. M.. “Job-shop scheduling with multi-purpose machines”. Silva. Doctoral Thesis on Production Engineering. integrating both fabrication of parts and their assembly into customer ordered products. . under today’s highly competitive markets it is important to provide good service to customers and. ”Job oriented production scheduling”.N. Inc. and Weglarz. Analysis and Development of Mechanisms and Algorithms to Support Product Oriented Scheduling. L.. W. at the same time. Ramos. G. B. Boston.. European Journal of Operacional Research. Blazewicks.H. C. “Analysis and Development of Mechanisms and Algorithms to Support Product Oriented Scheduling ”.. The main imbedded strategy used is what we described as Scheduling through JSP. Freund Publishing House. Baker. Almeida. (1967). “Scheduling in Computer and Manufacturing Systems”.J. The International Journal for Manufacturing Science & Production. “Sequencing and Scheduling”. Hastings. France. Carlos and Silva. Elmaghraby. Ana M. of course. pp 35-48. Ltd.R. (2000). C. J.J. described in detail. J.. Ecker. 8th ed. and Gambardella. S. We presented in this paper an approach oriented towards achieving very short manufacturing throughput times of customer orders. Heidelberg. 13 . Almeida. W. Maxwell. (2002). Brucker. Journal of Scheduling. (1998) “Production and Operations Management: A Life Cycle Approach”.

Silva. B. D. and Whybpark. L.. “Manufacturing planning and control systems”. Irwin. “Project on Intelligent Manufacturing cell at University of Minho”.C. 1st World Congress on Intelligent Manufacturing Processes & Systems. Puerto Rico. Inc. William. C. (1996).. 14 . S. T. (1995). E. Vollmann. Richard D. Mayguez.

Sign up to vote on this title
UsefulNot useful

Master Your Semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master Your Semester with a Special Offer from Scribd & The New York Times

Cancel anytime.