SE520367C2 - Method and device for interactive configuration - Google Patents

Abstract

A method for interactive configuration, in an interactive configuration system, by which a user may configure a configurable product by successively selecting i.e. including in the configuration, various items representing product parts, such that after user actions of this kind the accessibility of further items, is determined by means of an algorithm including a boolean SAT algorithm, that ascertains the satisfiability of formulas including those of propositional logic, representing the user choices, and said further items, together with a rule base of formulas stating relations of incompatibility and dependence characterized in that a user when wishing to select an item that is inaccessible due to earlier user choices is provided with a possibility of having a minimal number of earlier choices undone so as to retain a maximal subset of earlier choices compatible with the rule base and said inaccessible item, said undoing being in accordance with the user's relative preferences with regard to said earlier choices, by performing the following steps: (a) assigning an order of priority to the user's earlier choices, then (b) determining a maximal subset of earlier user choices with higher priority than all choices not included in said subset, said subset being compatible with the rule base and said inaccessible item, finally c) if the user accords a higher priority to said inaccessible item than to all previous choices excluded from said subset, forming a superset of said subset of user choices, being a maximal subset of user choices compatible with the rule base and said inaccessible item, formed in accordance with user priorities.

Description

40 45 55 60 65 5520 367 2 necessity.

The need for configuration arises at different levels. Of course, configuration solutions are required on the production side, i.e. in product design, but demand is also increasing on the consumption side. Consumers are increasingly demanding products that are tailored to their specific needs and tastes. With regard to special product segments (such as PC systems, cars and many other segments), there is thus a demand for computer-based systems for configuration management, which can be handled by a seller or the customer himself (eg a consumer who has to order a fi gureable product via the Internet).

Today, there are commercially available configuration management systems for both product design and consumption. These are based on various core technologies, such as extensive database search, programming of explicit constraints (constraint programming), batch logic SAT algorithms, etc. From a user perspective, there is a big difference between a system where you first make a series of choices and only afterwards find out if the choices correspond to a valid configuration or not, and a system that only presents the alternatives that together with already made choices form part of a valid configuration. The present invention is integrated into a system of the second kind using a Boolean SAT algorithm and it relates to a specific problem which may arise during the configuration process with the system in question. At some point in the configuration process, the user may find that certain choices they wish to make are not possible due to choices already made. In such a situation, there is no valid configuration that includes the choice the user wishes to make and all previously made choices. In this situation, it is desirable that the user be offered choices that are such that she can actually make the current choice while retaining a subset of previously made choices, which are compatible with the current choice and created based on her wishes.

Description of the invention The present invention solves a specific problem that arises in interactive configuration systems of a certain type. The program is also implemented as part of a specific system of this kind. Configurators of this type offer the user selectable modules comprising mutually exclusive alternatives, e.g. in the form of menus presenting mutually exclusive alternatives for certain aspects of the configurable product. In such a menu, there may be lists of colors of cars, different monitors for PC systems, etc., depending on which product is to be competed. Some options under a special menu may have been marked as unavailable by the user, because the user has previously made selections that are incompatible with the current selection. The user can select at most one of the remaining options as selected, which means that it must be included in the final configuration, or one / fl of the options that have been deselected, which in turn means that the option / one may not be included in the configuration. The configurator has a rule base (consisting of batch logic formulas) which contains (1) formulas that express incompatibility and dependency ratios etc. between the different options; (2) formulas that express what constitutes a complete configuration and (3) formulas that indicate which options are mutually exclusive. When the user selects an option that is selected, the atomic formula corresponding to the option is added to the rule base. If, on the other hand, an alternative is marked as deselected, the negation of the corresponding atomic formula is added to the rule base. The configurator then applies a Boolean SNF algorithm to check whether additional options are part of a valid configuration along with the options already marked as selected (selected). If this is not the case for a particular option, the configurator marks the option as unavailable.

To illustrate this, we can e.g. describe a configuration of a PC system using this type of controller - but as previously mentioned, the invention can be used in many other areas. For each selectable option under the menus _ and also for the menus themselves - there are corresponding atomic formulas. Atomic batch logic formulas are indivisible syntactic units that are used to indicate that something is the case. For example, if the user has selected the menu options hardware processor: Pentium III 550Mhz, hardwaremonitorzTrinitron 17 "and softwareazoszLinux selected, the configurator adds the corresponding atomic formulas pIII550, trinitron17 and linux to the rule base. Non-atomic formulas are built from atomic by means of connectives such as negation (n), conjunction (A), disjunction (V), material implication (->) and so on. These connectives express truth functions (Boolean functions). For example, the formula -windows has the value true if and only if the formula windows has the value false; the formula (pIII550 A trinitron17) has the value true if and only if both pIII550 110 115 120 125 130 135 140 5l2 0 .. 3.6 7 4 "/ and trinitron17 have the value true; the formula (pIII550 V trinitron17) has the value true if and only if at least one of the formulas pIII550 and trinitron17 has the value true, and finally the formula (pIII550 -> trinitron17) has the value true if and only if pIII550 has the value false or trinitronl? has the value true.

If the user has marked a menu item as deselected, the negation of the corresponding formula is added to the rule base. For example, if the user has selected the menu item hard disk hard disk: 1, 4GB as deselected, the formula -hd1.4 is added to the rule base.

In addition to the formulas that result from user choices, the rule base as we mentioned earlier contains: (1) formulas that express dependencies and incompatibilities between the selectable alternatives, such as the formulas ((linux / \ pIII550) -> -zip) which state that a computer system with a Pentium III 550 MHz CPU running the Linux operating system cannot be equipped with a zip drive, and (fi evmsof fi ce) -> (windows V macos) ) indicating that the Internet Explorer and Microsoft O fl ice programs require one of the operating systems Microsoft Windows or MacOS; (2) formulas that express what is a complete configuration, in this case a complete computer system, This is done using a formula of the following kind: (processor / \ memory / \ hard diskKeyboardA.. ./\Os), where the different conjunctions (processor, memory, hard disk, keyboard,..., os) are defined by formulas such as (processor - + (pII350 V pII400 V pIII400 V ... V pIII850)), (memory -> (32MB V 64MB V 128 MB V .. .V l024MB)), etc. Finally; (3) formulas indicating which alternatives are mutually exclusive such as (pII350 -> (fl pII400 / \ - \ pIII400 / \... / \ -Fi pIII85Û)) / \ (pII400 -> (Formulas of the second kind are required to: avoid situations where the user has made a number of choices that are compatible with each other but still do not form part of a complete configuration, ie the complete system requires a certain part that is not compatible with the user's choice. formulas is to determine whether all the formulas in the set in question can be true or not at the same time, ie whether they are satisfactory together.If we assume that the set in question contains adequate and sufficiently strong formulas of the type (1-3) above, in combination 145 150 155 160 165 170 175 520 367 with the formulas representing the user choices, determining whether the formulas in this set together are satisfactory essentially means essentially the same thing as determining whether the user choices are included in a guild tig configuration, i.e. in a possible (and accessible) combination of parts which together form a whole of the kind in question. A SAT algorithm accepts as input a set of formulas of the kind mentioned and determines whether the atomic formulas that occur in input formulas can be assigned truth values, so that all input formulas have the value true.

The invention specifically refers to a situation where the user during an interactive configuration by means of a configurator of the type described above, comes to a point where she wants to choose a particular alternative that has already been excluded due to her previously made choices, i.e. the option is marked as unavailable. For example, a user who configures a PC system at some stage may find that it is not possible to select the hardware memory memory24MB option. The reason for this may be that the user just at the beginning of the configuration chose the hardware processor: Pentium II 25OMI-Iz and there are no motherboards with enough memory space for 1024MB RAM, which are compatible with this type of processor and therefore the controller base base contains a limitation in this regard.

The user may much prefer to have l024MB of RAM than the selected type of processor. If she did not know that she could get 1024MB of RAM if she just chose another processor, she would probably like to do this and thus be much more satisfied with the configuration than otherwise.

In the situation described above, the user's current wishes (choosing a particular option) conflict with her previously made choices. Since the choice she now wishes to make is not compatible with (some of) her previous choices, she must deselect at least one (and probably fl era) of these previous choices in order to have her wish fulfilled. If she considers that the previous choices are more important than the current wish, she must of course accept that she cannot choose it. Otherwise, however, she might want to cancel conflicting choices that are of less importance to her so that the current wishes can instead be fulfilled. A method that offers the user this opportunity would add real value to her because in such a situation she would be able to use the program to configure a product that better suits her wishes than otherwise. Such a method would therefore be extremely valuable in the process of achieving an optimal configuration. The present invention relates to a method having exactly this ability.

Thus, if the user insists on making a choice that is not compatible with her previous choices, she is asked according to the present invention to rank her previous choices with respect to their relative importance to her.

Thereafter, the configurator deselects previously made selections in succession (however, they still remain in memory). The controller starts with the least priority selection and works its way up by checking in each step whether the remaining amount of user selection is compatible with the user's current wishes and finally stops when this is the case. This deselection is cumulative or additive in the sense that previously deselected alternatives' remain deselected during the process, and new deselected choices are gradually added to the set of deselected choices. In this way, the algorithm will detect a (lowest) point (in the order of priority), such that the set of the selected (selected) and deselected (deselected) alternatives above this point is compatible with the unavailable (unavailable) ) options that the user insists on choosing. An alternative, but logically equivalent method is to first deselect all previously made choices and then enter them from memory again in (descending) order of priority, cumulatively (ie they are added gradually to the amount of choices entered again) and then for each on this prioritize subset of the user's previous selection, check whether it is compatible with the unavailable (unavailable) option in question and terminate at the last (and largest) compatible amount.

Which of these methods should be chosen is a question of suitability _ both algorithms terminate at the same place in the list. However, one of them can be faster than the other, e.g. depending on the relative position of the termination point. Finally, the algorithm retrieves all of the user's previously selected options from memory and re-enters them. The user is now informed that if she gives the unavailable (unavailable) alternative in question a higher priority than the options excluded from the set according to this calculation, she can choose the option because it will then be compatible with the calculated set selected (and selected) and deselected (deselected) alternatives with a higher priority than the excluded ones.

If the user makes this re-prioritization and now tries to select the unavailable option in question, she will be told that this is possible but that the lower priority options (and deselected options) that she previously marked 215 220 225 230 235 240 245 I5201 367 ~ 7, and which are incompatible with the new option, in which case it will be deleted. If the user now selects the unavailable option in question, the lower priority options will first be deselected (ie no longer marked as either selected or deselected. However, the information about the user's previous selections associated with these parameters will remain in memory The algorithm then attempts to re-enter the previously selected options, one by one in descending order of priority, to obtain a maximum subset of previous user choices, which is compatible with the rule base and previously unavailable options.

A significant advantage of this method is that the user can choose an alternative that has been marked as unavailable due to previously made choices with which it is not compatible, while at the same time being given the opportunity to keep as many of the previous alternatives as possible. . In this way, the user can obtain a product that conforms to her wishes and with which she can feel much more satisfied - in a shorter time and with lower work effort than otherwise.

Prior Art 7099/13411 discloses a prior art relating to a problem of the same kind which can be solved by the present invention. The method in question is intended to be integrated into an interactive configurator of the type described in VVO99 / l3413 by the same patent applicant.

When a user of this configurator wants to select an selectable option, which due to previous selections has already been marked as unavailable, the system will identify all maximum amounts of previous user selections that are compatible with the unavailable option in question as well as the rule base.

A list of maximum quantities is now presented to the user who is asked to select one of them. Such a list of quantities can of course be quite long in some cases and also difficult to overview. To facilitate the selection, the user is given the opportunity to select some of the previous choices that must be included in the maximum quantities presented to her. In this way, the number of options that the user may have can be reduced.

According to this prior art method, maximum amounts of previously made choices are identified by going through all alternative ways of deselecting one of the previous choices and then all alternative ways of deselecting two previous choices and so on. Each time it is checked whether the amount of remaining choices 250 255 260 265 270 275 280 1520, 367 is compatible with the unavailable alternative in question and the rule base. If we let P = {p1,. . . pk} represent all (k) previously made choices, and q represent the unavailable option that the user insists on making, the method first checks whether any of the sets P \ {p1}, P \ {p2},. . . P \ {pk} is compatible with q and the rule base. Then it goes on to check if any of the sets P \ {p1, p2}, P \ {p1, p3}, .. .P \ {pk_1, pk} are compatible and so on. In this way, all maximum amounts of P that are compatible with the rule base and q are identified.

A major disadvantage of this prior art, compared to the present invention, is that the number of compatibility checks is much larger, which results in significantly more time consuming calculations and, of course, reduced user-friendliness. According to the present invention, only one (1) method is required to cancel a user selection in P = {p1,. . .pk}, while in the method described in WO99 / l3411 there are k ways to consider when canceling a selection. Furthermore, only one (1) way is required to cancel two choices instead of and usually only 1 way to cancel n choices instead of different ways.

Consequently, the largest number of repetitions, according to the present invention, becomes equal to the number of previously made user choices minus one (since the empty set Q) does not need to be checked), while the number in the method described in \ vo99 / 1s411 now amounts to gig, _ 1 = zk _ 2, where k is the number of previously made user selections.

Intended implementation A suitable implementation of the invention is illustrated by means of a number of flow charts (fi g. 1-4) which are commented on below.

Flowcharts describe algorithms that are preferably implemented in a computer program. It must be possible to run it in a monitor-connected computer that is connected to some input devices, such as keyboard and mouse. Using this program, a menu system as described above will be displayed on the monitor. Using the input devices, the user can mark a menu item on the monitor as selected or deselected. Some options can also be marked as e.g. not available due to program activity. Other input and output events are also exchanged via these input and output devices between the user and the computer in which the program is running. The computer running the program may be set up in one place and connected to the monitor and input devices - directly 285 290 295 300 305 315 -520- 367. 9 or indirectly - via a computer located elsewhere. It may also be the case that parts of said program are executed in a computer and another part in another computer connected to the first.

To facilitate the following description, we start by defining a few terms: The menu options that the user can mark as selected or deselected and that the program can mark as unavailable, we call selectable options. In the said intended implementation, each selectable alternative corresponds to a variable that represents this. We call these variables selectable variables. The selectable variables relate to the selectable alternatives in such a way that they assume the value 1 when the corresponding selectable alternatives are marked as selected, and 0 when the corresponding selectable alternatives are marked as deselected. Furthermore, each selectable variable corresponds to an atomic theorem formula that expresses the value of the variable in question. If p refers to a selectable variable, we use p to refer to the corresponding atomic formula. For example, that the value of p is 1 is expressed by the formula p, and that it is 0 by the negation -p of the formula p. If the variable on the other hand is not finiered - which is the case when the corresponding menu item 'is not highlighted ~ fi no formula expresses this. The absence of both formula p and - = p thus indicates that the corresponding variable p is not defined and that the corresponding menu item is not selected.

We refer to the selectable alternative that corresponds to the selectable variable p with p. By virtue of 'these equivalents, we will in the future for the sake of simplicity talk about quantities of selectable variables that are compatible or incompatible with the rule base. What is meant, of course, is that the set of formulas that express the value of these variables is either compatible or incompatible with the rule base.

In the intended implementation of the invention, the program comes after each user input, ie. after the user has selected a menu item that is either selected or deselected, to go through the menu items that the user has not yet selected. For each of the alternatives q, the corresponding formula q is added to the formula set corresponding to the set p1. . .pk of selectable variables that represent previous user selections. The set will then be sent together with a set of formulas that express different constraints for the selectable alternatives (rule base) to an implementation of a SAT algorithm to determine if the set is consistent.

If not, the menu option (q corresponding to the variable 320 325 335 340 345 350 355 520 "367/0 q will be marked as unavailable. If the quantity on the other hand is consistent, the next unselected option will be checked on the same way.

We call an option marked as unavailable simply unavailable menu item and generally talk about the availability of such a menu item and then refer to whether it is available or not, as appropriate. The amount pl. . . pk of previous user selections are assigned a default order of priority, which is based, for example, on the order in which the user marked the options in question, or on any user profile that relates to the specific user, or on some other principle.

The flow chart in Figure 1 describes the following possible implementation in accordance with the invention: The algorithm is activated when a user wishes to select an optional alternative F; which is marked as not available due to previously made user selections represented by the selectable variables pl. . pk. To begin with, the variables are assigned a default order of priority in which pl has the highest and pk the lowest priority. For 1 g n, m g k if m <n then pm is assumed to have a higher priority than pm i.e. there are no alternatives that have the same degree of priority. First, the user is asked if she wants to rank her previous choices based on how important they are to her, i.e. make a re-prioritization. If she answers No, the algorithm will apply the pre-specified priority order and directly call subroutine S2 (fi g. 3) which identifies a set of P selectable variables that represent previous choices that are compatible with q and the rule base. This subset of {p1. . .pk} is maximized and prioritized in a way that will be outlined below (fi g. 3). If the user instead answers Yes, the algorithm first enters subroutine S1 (fi g. 2), which enables her to re-prioritize her previous choices pl. . pk, to then proceed to subroutine S2. Since PU {q} is compatible with the rule base, the choice is in accordance with the user's wishes. This is provided that q is assigned a priority level that is at least as high as for any of the options in P, or in other words higher than for all options in {p1. .pk} \ P (see explanation of lig. 3) The user is now asked if (1 has a higher priority for her than all the options corresponding to {p1 _. .pk} \ P. If the user answers No, the algorithm then ends the choice of the alternative G1 (which corresponds to variable q) is obviously not compatible with the user's wishes, which she is also informed about.If the user on the other hand answers Yes, the algorithm enters subroutine S3 (fi g. 4) which identifies 360 365 370 375 380 385 390 = s2o 567 ll an amount Q so that PQQ C_l {pl. .Pk}, QU {q} is compatible with the rule base, and Q is maximum in a way that we will explain below (fi g. 4) .

The flow chart in fi g. 2 describes subroutine S1 of the algorithm represented in fi g. 1. The subroutine begins with a plurality of pl. . .pk of selectable variables that represent previous user choices ranked in descending order of priority as above (fi g. 1). First, the user is asked to select an option from the list pl. . .pk to re-prioritize it, i.e. an alternative that she wants to assign a different, relative priority in relation to the other alternatives in the list. The user selects such an option pm (we call this variable s).

Remaining variables pl. . . pm_l, pmil. . .pk keeps its respective places in order. We will henceforth call them pl. . .pk_l (for m <n § k, pk now gets the name p ,, _ l).

At this stage, the user is asked to select a new priority level (one of for the selected option s, whereupon she selects a new priority level n 6 {1.... Then the algorithm arranges the variables pl. _ .Pk_l and s as follows: Förj <n retains respectively pj its place in the list; s assigned position n and for n í j <k, assigned pJ position j + 1.

The user is now asked if she is satisfied with the new priority scheme pl. . .pk for the variables or if she wants to make further reprioritisations.

If the user selects Yes, the algorithm enters the subroutine again. If she chooses No instead, the algorithm goes out of this subroutine. The subroutine again produces a list of user choices pl. . .pk which are now ranked based on user preferences. The subroutine can of course be implemented by standard routines for manipulating ordered data structures.

The flow chart in fi g. Fig. 3 describes subroutine S2 of the algorithm shown in Fig. 1. Just as subroutine S1, the subroutine begins with an amount {pl. . .pk} variables that represent previous user selections, ranked in descending order of priority, and a variable q that represents any unavailable selectable option that the user insists on selecting. First, the subroutine initializes a counter l to the location of the last option in the set {pl. _ .pk}, i.e. l is assigned the value k. Then atomic formulas and negations of atomic formulas corresponding to the variables p1 are added. . .pl_l and q = 1 to the rule base (q and for 1 gm 3 l - 1 either pm or ppm depending on whether the value for pm is 1 or 0), and a SAT algorithm is called to check whether this set of formulas is satisfactory . If it is not, the subroutine first checks for the amount {pl. . .pl_l} is empty ({pl... pll} is the empty 395 400 405 410 415 420 425 52 0.. 3.67 / - '¿set (Û). In that case q is not compatible with the rule base which means that it it is not possible to select the selectable option q unless the rule base is changed, therefore the algorithm terminates in this case. On the other hand, if the set (pl... p¿_1} is not empty, 1 is removed from the counter l and the subroutine continues to con- conjure the subset {p1.. .p, _2} of {p1.. .pk} in the same way.

In each repetition, the subroutine in this way removes the lowest-priority alternative, which is why the removal in this respect becomes (reverse) successive.

The deletion is also cumulative in the sense that deleted alternatives remain deleted and new deleted alternatives are added to the amount of previously deleted alternatives with each repetition.

If q is compatible with the rule base, then for some l> 0 there will be a subset (pl.. .P¿_1} of {p1.. .Pk} that is compatible with q and the rule base.

The subroutine will find a subset of P that is structured in such a way that each element in P (if any) has a higher priority than all elements in {p1 ... pk} \ P. Thus, this subset is an initial segment of {p1 .. .pk}. The amount P is prioritized in this respect. P = {p1. . .p¿_1} is also maximum in the sense that if l '> l then {p1 ... p¿f_1} is not compatible with q and the rule base. In this way, P is the maximum initial segment of {p1 ... pk}, which is compatible with q and the rule base. When the set P is identified in this way, the subroutine ends with q, {p1 ... pk} being returned and l indicating the element p, i {p1 ... pk} which has the highest priority outside the set P.

The flow chart in fi g. 4 describes subroutine S3 for the algorithm represented in fi g. Subroutine S3 starts with the parameters q, (111... Pk} and l which are transferred from subroutine S2 and which represent unavailable optional alternative q that the user wishes to select as well as the maximum and priority (in the manner described in fi gur the explanation above) the subset P that is compatible with q and the rule base The subroutine has the task of identifying a maximum subset R of {p1 ... pk} \ P so that Q = RUP becomes compatible with q and the rule base and is generated taking into account to the relative priority that the user assigns the elements in the set {p1... p ,,} \ P as follows: Let RB be the rule base. We say that a subset S of (191... pk} is maxi'- malt q- consistent (with respect to RB) if SU {q} U RB is a maximum consistent amount relative to {p1 ... pk}, ie if SU {q} U RB is consistent and if r E {p1... pk } \ S then SU {r} U {q} U RB is not consistent.

Q is obviously a maximum q-consistent extension (superset) of P in 430 435 440 445 450 520 367/3 in this respect. If Q 'is any other maximum q-consistent extension of p, it means that any element in r in Q \ Q' has a higher priority than all elements in Q '\ Q. Thus, Q is generated in a manner that better conforms to the user's desires than all alternative maximum q-consistent amounts.

First, a counter variable n tilll is initialized. Then the formula set corresponding to the variables {p1 _ is added. _ p ,,} and the formula q to the rule base and then the formula set is sent to the SAT algorithm. If the set of formulas is not satisfactory (which of course is the case if n = 1), the value of the variable p ”will be set to undefined, which corresponds to a delimitation of the corresponding selectable alternatives. If, on the other hand, the set of formulas is satisfactory, the value of p 'remains unchanged. Then it is checked whether n <k and if not the end of the subroutine. Otherwise, the counter n increases incrementally and the subroutine begins a new iteration. If {p1. . .p ,,} now contains undefined variables, we must be clear that no formulas corresponding to these variables are included among the formulas corresponding to the variables in this set, because in that case ~ as we previously pointed out ~ there are no corresponding formulas .

The subroutine thus works in a .successful way in the sense that it goes through the list of alternatives. . .pk in order. In addition, it is cumulative or addítw in that the alternatives that have been added during previous preparations the amount whose compatibility is to be checked, remains despite the addition of new alternatives.

Claims (3)

520 "367 f? Patent claim Method of interactive configuration in an interactive configuration system by means of which a user can configure a configurable product by successively selecting, i.e. include different options (alternatives) that represent different product parts of the configuration, such that the availability of additional options is determined using an algorithm that includes a Boolean SAT algorithm that determines the satiability of formulas, including batch logic ones, which represent the user choices mentioned above and above. additional choices together with a rule base consisting of formulas expressing incompatibilities and dependencies, characterized in that a user who wishes to choose an alternative that is not available due to previous user choices, is given the opportunity to cancel a minimum number of previous choices in order to thus being able to maintain a maximum subset of previous choices, which is compatible with the rule base and the unavailable alternative. The cancellation is performed in accordance with the user's relative preferences with regard to previously made choices by performing the following steps: a) previous user selections are assigned a priority order, then b) a maximum subset of previous user selections with higher priority is determined than all the selections not included in said subset, such that this subset is compatible with the rule base and said unavailable alternatives; and finally, c) if the user assigns said unavailable alternative a higher priority than all previous selections excluded from said subset, then an extension of said subset of user selections is generated, which is a maximum subset of user selections compatible with the rule base and said unavailable options, created based on the user's wishes. Method according to claim 1, characterized in that the access to further alternatives is determined after each user entry of said kind. Method according to claim 1, characterized in that the user can also deselect, i.e. exclude, different options that represent product parts from the configuration. 35 40 45 55 60 10. 11. 520 367 IS '. Method according to claim 1, characterized in that the rule base contains formulas which express requirements for completeness. Method according to claim 1, characterized in that the subset determined in step (b) is achieved by successively and cumulatively canceling said previously made choices, in ascending order of priority and in each repetition checking with said algorithm whether the amount of remaining user choices is compatible with the rule base and the said unavailable alternatives. Method according to claim 1, characterized in that the subset determined in step (b) is achieved by successively and cumulatively creating a subset of said previous selections, in descending order of priority and in each repetition checking with said algorithm whether this subset of user selection is compatible with the rule base and the said unavailable options. . Method according to claim 1, characterized by! that the subset created in step (c) is obtained by successively and cumulatively, in descending order of priority, adding said excluded alternative to the subset determined in step (b), and in each repetition checking with said algorithm whether the extension of said subset created in this way is compatible with the rule base and said unavailable alternatives, and if so, retain the excluded alternative, and if not exclude it. . Method according to claim 1, characterized in that the order of priority assigned to previous user selections in step (a) is performed by the user himself. . Method according to claim 1, characterized in that previous user selection in step (a) is assigned a default order of priority. Method according to claim 9, characterized in that the user can subsequently modify the preset priority order. Method according to claim 9, characterized in! of the preset order of priority being determined on the basis of the order in which the user made his previous choices. 70 75 80 90 95 100 100 13. 13. 520 367/6 Method according to claim 9, characterized in! that the preset order of priority is determined on the basis of a user profile relating to the user in question. Device comprising a terminal with monitor and input units, connected to a computer in which a program is run which via the monitor presents a menu system to the user so that it can use the input units to perform an interactive configuration of a product by successively selecting, i.e. in the configuration include various menu options that represent product parts and where the program implements a method that after each user input makes the availability of additional menu options determined using an algorithm that includes a Boolean SAT algorithm that determines the satiability of such formulas ~ including batch logic w which represents the user's choice and additional menu options, together with a rule base with formulas expressing incompatibilities and dependencies and where access to additional menu options is presented to the user via the monitor characterized by a user wishing to select one due to previous user selection not available alternative is given the opportunity to cancel a minimum number of previous user choices in order to be able to maintain a maximum subset of previous choices, which is compatible with the rule base and the said unavailable alternatives. The cancellation is made on the basis of the user's relative preferences with regard to previously made choices by performing the following steps: a) previous user choices are assigned a priority order, then b) a maximum subset of previous user choices with higher priority is determined than all of the choices not included said subset, such that this subset is compatible with the rule base and said unavailable alternatives; and finally, c) if the user assigns said unavailable options a higher priority than all previous selections excluded from said subset, then an extension of said subset of user selections is generated which is a maximum subset of user selections compatible with the rule base and said unavailable options, created in accordance with the user's wishes.