Keywords: case-based reasoning, constraints satisfaction problem, ceiling voids layout, artificial intelligence, complex geometry, large problem.

In this paper an object-based CAD program is used to take advantage of standardization to handle the schematic design, sizing, layout for services in a building ceiling void. From the specification of the building 3D model, our software proceeds through different steps; from the determination of the standard number and size of fan coils to the generation of 3D solutions. In order to deal with more complex geometry and larger problems, we have used a hybrid approach: Case Based Reasoning (CBR) [1] with the Constraint Satisfaction Problem (CSP) [2] approaches. In practice, engineers in building services use previous solutions and adapt them to new problems. CBR mirrors this practical approach and does help us to deal with increasingly complex geometry effectively and meanwhile the CSP approach has been used for layout adaptation.

CBR is an approach able to utilise the specific knowledge of previously experienced, concrete problem situations named cases. Through recalling these cases and reasoning with them, solutions to similar situations can be found. A new problem is solved by retrieving a similar past case, and adapting it in the new problem situation. CSP is a powerful and extensively used artificial intelligence paradigm. It is a natural way of representing problems because the user needs only to state the constraints of the system to be modelled. A CSP is defined by a set of variables, a domain of values for each variable, and a set of constraints between each pair of variables. A solution of the CSP is a consistent assignation of all variables to values in such a way that all the constraints are satisfied. The benefits of integrating CBR with CSP include formalization of the CSP and to make the CBR processes of adaptation, retrieval, matching, etc. domain-independent. CBR can help CSP when models are incomplete or the domain knowledge is sparse. CSP can also help manage the problem complexity via constraint propagation, weighted constraints, etc. CBR can help increase CSP efficiency by reusing past experience, which also aids in knowledge transfer and in the consistency of approach.

The ceiling void solution is based on a four-duct-fan-coil system. Fresh air is provided by a central air handling unit. Air supply to the space is via slot diffusers in the perimeter zones and the square diffusers internally. The non-service elements for ceiling voids layout include beams, ceiling tiles and core areas. In order to retrieve the similar case from the case base, non-service elements and floor zones need to be defined at first. After the similar case has been retrieved, the associated logical layout case is then activated to layout the equipment (fan coils and diffusers).

To adapt the previous case to meet the new problem, a series of constraints are used to adapt the case, they include dimensional constraints, inclusion constraints and non-overlapping constraints. The dimensional constraints assign the minimal or maximal values to the object constrained variables, and the constraint is expressed by equality or inequality. The inclusion constraints ensure the objects layout within the defined space and are not placed outside the space boundaries. The non-overlapping constraints are represented as the fact that two objects do not overlap with each other. For equipment layout, the constraints are used to ensure that the equipment are laid out within their local grids (each grid can hold one fan coil and the assigned diffusers to service a typical unit area). At same time, the equipments must non-overlap with each other, ducts and non-service structures nearby. For ducts (local ducts and distributed ducts) layouts, the typical constraints are used to create the duct route. The duct route must have the shortest length, minimizing the length of each duct using the "branch and bound" algorithm; the joints number (bends) of duct must be minimized; Every two neighbouring duct segments must have different directions, and their bended angle must be greater than 90^o. Duct routes must not overlap with equipment and non-service structures (i.e. columns, stairs and lifts etc).

To conclude, standard solutions in conjunction with the IT have the potential to significantly reduced design costs by reducing the design time which is estimated to be typically in the region of 70%, and improve the quality and produce additional benefits elsewhere in the supply chain. The idea to have a compromise between full automation and interactivity gives the designer full control of the design while assisting him to solve complex problems automatically. This compromise is the main difference with the aforementioned approaches in facilities layout and duct/pipe routing.

1

A. Aamodt, E. Plaza, "Case-Based Reasoning: Foundational Issues, Methodological Variations, and System Approaches", AI Com - Artificial Intelligence Communications, Vol. 7, No. 1, p.39-59, 1994.

2

C. Baykan, M. Fox, "Constraint Satisfaction Techniques for Spatial Planning", Intelligent CAD Systems III, Practical Experience and Evaluation, 1991.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £80 +P&P)