Cib 2004 conference: tuesday morning¶

On monday night I went to dinner with an Saudi-Arabian, a Swede and a Brit :-) We went to a jazz restaurant where they had a band playing for a fund raising activity, the pianist was said to be quite famous: Norm Amadio. Great music!

This morning there were a couple of good talks. Read on. Hennes de Ridder (TU delft): process and system innovation in the building industry

Recenty Hennes, in cooperation with a few others, developed a simple model that works quite well. It's the value-price-cost model. The consumer wants value (expressed in money). He has to pay a price for a good (money). And a producer has cost associated with creating the good (again money).

The added value for the customer is the value minus the price, the profit for the producer is the price minus the costs. The total benefit is the added value plus the profit.

Now, the bigger the gap between costs and value, the bigger the cake that can be split into added value and profit. When looking at it this way, it is important for both parties to make the room between costs and value as big as possible.

In reality the demand side is fixed in the demand specification. At that point the supply side is mobilised. So there is no interaction, no possibility to get a really attractive solution. Everything is fixed, though there is a lot of uncertainty... Which results in war during the project and claims after the project. Which means extra costs and less value, making the cake smaller...

Adequate project control must be dynamic. you must be able to accept a bit less value if that brings the cost down a lot. likewise, you can accept a bit more costs if that raises your value a lot. Just common sense.

The link between client value and price is the client's brief, between producer cost and the price it is the design. So those two items and the link between them must be dynamic. Continuously matched to eachother, step after step.

Conclusions

• Make a difference between non-negotiable requirements and measurable performance.
• Measurable performance should be the basis for the contract, instead of pre-defined output specifications. As an example, take the Rotterdam storm surge barrier. In practice, they calculated all the bad things that could happen and came up with a design that could handle all those risks. As an effect, the structure is basically over-dimensioned. In this newer theory, they would have set a 1-per-10000-years max failure rate as non-negotiable. The measurable performance would have been the yearly difference between the sealevel and the river level, as that determines the load on the structure. The solution would have been in the hands of the constructor (and in this scenario, therefore the maintainer!) and he could have used a lighter structure. If the trend for the sea level would have started to rise, they could make the structure more heavy, if needed.

  In a normal scenario this would have meant an extra contract, which means a lot of money. So in the normal scenario you just design it to be heavy from the beginning.

• The idea is to have a dynamic control of value/price/cost, keeping it in balance throughout the lifecycle. (This assumes a design-build like contract).
• Involving all stakeholders (demand and supply) in a life cycle value alliance, and committed to maximise total benefit. (Thus making the pie bigger for all and thus the individual shares).

Answers to questions

• Process risk is for the supplier, requirement risk is for the customer.
• In the current building practice, a contractor loves extra work, as it means more money. In the value/price/cost model, he loves to lower costs, as it widens the gap between value and cost, so more money!

Scott Fernie (Reading uni)

Practice and knowledge is always embedded in a context. The research started with a report on knowledge exchange between the aerospace context and the construction context (which i might write a litte on later, as Scott promised to send me the report).

In the tender and selection process there are huge costs. For example, you need to do a lot of expensive work to absorb the offered job into a more technical design and, especially, a price which won't gut you but which is low enough to get the job. Which you might not get, therefore losing you the money you spend in the process. There is also a lot of friction costs in this normal building process, here called the normal context.

In the Uk there sometimes is a bit more trust in cases where a customer wants to work with a limited set of contractors. They get a continuity of workload and the customer gets the certainty he won't be cheated. In such a context, you need to trust eachother. you need to learn to trust eachother. It is still in the early stages, but the context seems to be emerging.

He calls it the integrated procurement context.

(Note: when writing it down, part of the story seems to be missing, which is because the talk was too interesting to write everything down :-)

Maurits Dekker (TNO): iBuild

iBuild is intended for mass customisation for housing projects. some big points:

• The customer configures his own house beforehand.
• The customer walks through his house in VR.
• The building process itself is industrial.

There are 4 configuration levels;

• Neighbourhood.
• House.
• System (i.e. front of the house for example)
• Object (for instance a door)

The focus is on:

• Consumer driven house design.
• Use of international standards for good communication (IFC, owl)

Farnaz Sadeghpour (Concordia uni): geometric reasoning for site layout

There is a distinction between site objects, the fixed things like entrances, existing buildings, and construction objects, like material, equipment. The last category is what needs to be placed by the spatial analysis model.

It is implemented in Autocad, VBA and msAccess.

According to (More, 1980), there are basically two methods to do this:

1. Place everything, in a couple of combinations, and try to improve it that way
2. Bring objects in one by one, calculating a new optimal placement after each step.

The second one was the method used. In this method, the order of placement is important. The queuing criteria can be calculated by the program by entering values and weight factors for for instance cost of relocation, area size, lenght of time the object is on-site, etc.

The constraints are also weighted and fitted out with a score on how much they contribute to the degree of satisfaction with the resulting solution. The calculation could be presented graphically (colorcoded, the greener the area, the better fitted for placement), as it was implemented in Autocad.

The current implementation only handles direct-line distances and rectalinear (first in x-, then in y-direction) distances.