Business issue to enable line balancing :
Building a valid, manageable Precedence Graph
Business issue to enable line balancing :
The inability to effectively gather, validate and manage the precedence graph of all constraints for the product (the car) and the technology (the assembly line) then becomes a top business issue ! Given its high complexity, many automotive manufacturers currently still operate with on limited knowledge of the applicable product and/or technology constraints, have only limited precedence graphs available, multiple graphs from various departments conflict with each other, and sometimes even conflict with the actual manufacturing sequences of processes.
Clearly, solving this business issue must become a management priority to succeed in improving assembly plant performance. This overall business issue can actually be seen as three separate business issues adding to each other: how to build the precedence graph for a product, how to guarantee its reliability, how to keep managing the graph as the product evolves, as products are added [!!!],.
Let's look at these three steps in detail now.
1. How building the precedence graph ?
Let's examine this issue, from the perspective of the typical questions that would be raised in real business situations.
I have several precedence graphs (from product design, engineering planning, assembly shop floor, quality control), but they are different and even seem to conflict with each other. How can reconciliate them into one that I can trust and can be used everywhere ?
Yes various precedence graphs from various sources, partial or not, can be reconciliated.
Conflicting precedences can be identified with the appropriate tools. PrecBuilder claims to be the best available tool n this specific and complex task.
Currently my precedence graph contains uncertainty (it is not complete) ? Why is this a risk ? What dangers am I heading for ?
Leaving unanswered relationships between operations leaves the risk that illegal operation sequences could be suggested at optimisation stage. That's why it is absolutely essential to build a complete (fully validated) precedence graph. As long as the graph...
How much effort will still be left, when I would have answered the top recommended question to solve by PrecBuilder ?
As a result of PrecBuilder's unique Power-analysis feature, a list of most recommended relationships to clarify (“top questions”) will be offered for investigation. Next to each of these question a residue count information is displayed, showing how many questions to answer would be left.
What if the most recommended question to solve seems pretty unlikely to be confirmed, did I lose my time in looking at it ?
No, no time was lost. Knowing that something is NOT true, is yet adding knowledge. Therefore, the knowledge of the general process being looked at will still be enriched. Also, the top recommended list of question allows to easily select another narrow question to answer : it's resolution power might be smaller (i.e. its residue count might be higher) but still, the information is valuable and will help in validating the overall graph.
2. How guaranteeing the reliability of the precedence graph ?
Let's again examine this issue, from the perspective of the typical questions that would be raised in real business situations.
When I update, manage, maintain my precedence graphs with changes (adding constraints, adding validation information from assembly, etc.), how can I be sure the resulting graph will still work (i.e. be complete and feasible) ?
As each information is added, like “operation A must precede operation B”, an immediate check is done across all graph to ensure that no conflict is created (e.g. that some earlier information stated that “Operation B must precede Operation A”) and that no inconsistencies are introduced (e.g. that a cycle is being introduced). PrecBuilder uses the advanced mathematical algorithms of its resolution engine to ensure that at each moment in time the current graph is feasible (i.e. compliant with all constraints and intelligence information provided).
In a large scale and multi-user environment how is the reliability still guaranteed ?
The design of PrecBuilder and its data model allows to apply its validation and power-analysis features as well on partial as on complete graphs. As a result editing the graphs and reconciliating them even in an asynchronous, distributed, multi-user environment remains possible, and one of the strongest strength of PrecBuilder.
How does Optimal Design's PrecBuilder actually work ?
PrecBuilder is a powerful precedence graph builder that relies on advanced mathematical algorithms (graph theories, optimization technologies) to guarantee the feasibility of the precedence graph being built, but also to strongly assist the end-users in an efficient and effective graph building in a collaborative environment.
PrecBuilder takes into consideration existing constraints information from many sources and all available intelligence information from the assembly line across various product mixes. While collecting and merging the information, PrecBuilder suggests the most interesting precedence relationships to solve, immediately detects precedence conflicts and traces back to all information sources justifying each precedence information for an easier resolution.
3. How managing guaranteed reliable precedence graphs ?
Let's again examine this issue, from the perspective of the typical questions that would be raised in real business situations.
As the precedence graphs evolve, how can I revert some changes or identify who applied what changes and what were their consequences ?
Yes, if you're using PrecBuilder each input and graph enrichment step is traced in time and the whole graph evolution is seen as a multi-product evolving project which can own a complete history at both product and project levels. As a result, each relationship between two operations in the precedence graph can be fully drilled down and PrecBuilder will display the list of all mathematical justification (and those who stated them) that led to the examined relationship. If necessary, a previous version of the precedence graph can be restored and analysed.
Once my precedence graph is complete, what else do I need to securely optimise my assembly line ?
Once the precedence graph of constraints is complete, all products constraints are supposedly known and have been incorporated, as per the product expert's knowledge. In order to optimise the assembly line, use a tool like OptiLine to feed the precedence constraints graph, take technological constraints into consideration and automate this line balancing process to identify the optimal compatible assembly sequence.
What are the risks that will remain in building a precedence graph ?
Of course over-constraining a graph might significantly reduce the future optimisation potential. For instance if “Operation A may follow Operation B” is turned into “Operation A must follow Operation B”, the relationship between these two operations will be over-constrained. The deployment and implementation approach proposed by Optimal Design takes care of this risk and mitigates it thanks to its methodology.
|
