Pemilihan Tipe Komponen D 0394 Perancangan Sistem Manufaktur

Pemilihan Tipe Komponen D 0394 Perancangan Sistem Manufaktur Kuliah Ke XIX - XX

Flexible Manufacturing Systems (FMS) • An FMS is a “reprogrammable” manufacturing system capable of producing a variety of products automatically. • Conventional manufacturing systems have been marked by one of two distinct features: – The capability of producing a variety of different product types, but at a high cost (e. g. , job shops). – The capability of producing large volumes of a product at a lower cost, but very inflexible in terms of the product types which can be produced (e. g. , transfer lines). • An FMS is designed to provide both of these features.

FMS Component • Numerical Control (NC) machine tools • Automated material handling system (AMHS) – Automated guided vehicles (AGV) – Conveyors – Automated storage and retrieval systems (AS/RS) • Industrial Robots • Control Software

Texas A&M Computer Aided Manufacturing Lab Pratt & Whitney (MP 1) Leadwell Adept (MP 2) (MH 1) Port (P 1) Sabre 500 (MP 3) GE-A 4 Robot (MH 3) Manual Wksstn 1 (OP 1) Manual Wksstn 2 (OP 1) Puma Robot (MH 2) Port (P 3) Kardex (ASI)

Classification of FMS-related Problems • Strategic analysis and economic justification, which provides long-range, strategic business plans. • Facility design, in which strategic business plans are coalesced into a specific facility design to accomplish long -term managerial objectives. • Intermediate-range planning, which encompasses decisions related to master production scheduling and deals with a planning horizon from several days to several months in duration. • Dynamic operations planning, which is concerned with the dynamic, minute-to-minute operations of FMS.

FMS Problems • Part type selection (Askin) - selecting parts that will be produced in the FMS over some relatively long planning horizon. • Part selection (Stecke) - from the set of parts that have current production requirements and have been selected for processing in the FMS, select a subset for immediate and simultaneous processing. • Machine grouping (Stecke) – partition machines into groups where each machine in a group can perform the same set of operations. • Loading (Stecke) - allocate the operations and required tools of the selected part types among the machine groups. • Control - provide instructions for, and monitor the equipment in the FMS so that the production goals identified by the above problems are met.

FMS Part Type Selection Problem • The part type selection problem is involves determining which parts should be produced in an FMS. – In theory, there should by some savings achieved by producing some parts in the FMS rather than producing them on manual equipment or purchasing them from outside. – The objective is to load the FMS to maximize the savings achieved by producing parts in the FMS subject to FMS capacity constraints. • Typically, the part type selection problem is solved infrequently For example, the decision to outsource a part rather than producing it in-house requires time for the subcontractors to do the required design and tooling.

Part Type Selection Formulation • Parameters/Variables: • P is the production time period available on the key, bottleneck resource. • pi is the time period (time per unit ? units period) required by part type i on the key bottleneck resource. • si is the savings period (savings per unit ? units period) if part type i is added to the FMS. • xi is the binary decision variable indicating whether or not product i is produced in the FMS.

Part Type Selection Formulation • Knapsack problem – IP with a single ‘capacity’ constraint – LP relaxation can be solved by inspection. – Solution method : • Heuristic • Dynamic Programming • Branch and Bound

Greedy Heuristic for Part Type Selection 1. Order the part types [1] to [N] such that: 2. For i = 1 to N, select part type [i] if si > 0 and inclusion is feasible. 3. Example: Given the following parts, and assuming that there are 250 hours available on the FMS, use the above heuristic to perform the part type selection:

Example (cont’d) Parts Unit Savings Unit Proc. Time Demand si 1 50 1 100 5000 100 50 2 100 1 50 5000 50 100 3 70 . 3 75 5250 22. 5 233. 3 4 30 . 5 300 9000 150 60 5 25 . 2 25 625 5 125 Sequence: 3 -5 -2 -4 -1 Assignment: Part Hrs. 3 22. 5 5 5 2 50 4 150 pi si/pi Cumlative 22. 5 27. 5 77. 5 227. 5

Optimal Solutions • Dynamic programming – Principle of Optimality - For any initial stage, state, and decision, subsequent decisions must be optimal for the remainder of the problem that results from the initial decision. In other words, given a stage and state, the optimal decision from that point on is independent of how you got to that state. – The decision of whether or not to include each part type in the FMS is a stage. The state is the remaining processing time on the FMS.

Dynamic Programming • Let fi( ) be the cost savings for the optimal decision regarding part types 1 to i if they occupy time units period on the FMS. If s 0 For 2 i N

Comment on the Part Type Selection Problem • Assumes that either all or none of a product is produced in the FMS. • Assumes that there is a single key, bottleneck resource and that the other resources have sufficient capacity to handle any configuration of parts • Assumes that the pi’s are independent of the parts selected. • The si’s can be very difficult to assess