Introduction: 1. level

(Public) Procurement

The statement from the General Accounting Office provides an unambiguous situation report. But the codification processes and the transforming of social reflexes may be long and complex. The first presumes the necessary majority in parliament. The latter should be basically only wanted.
The MY-X FREE services (c.f. COCO, COCS, COCO-STEP for evaluating variations and STEP IX for deriving real price intervals) would like to support the transaction of tendering (as soon as possible automated) by responsible experts in firms/institutions (starting public procurement processes), applicants, participants of supervisory boards, and their background experts.
The MY-X FREE services provide robust answers automatically in advance in case of following questions: which offer should be declared as winner based on all potential or real offers? How the evaluating process can be made more efficient (c.f. saving costs, time and avoiding conflicts and suspicions)? Which argumentation can be accepted as most standard in judicial processes?
The public procurement processes can be supported, inasmuch as the legitimate data-assets (collected and permitted for public using by the necessary authorities) are given for each relevant objects and for their relevant attributes.
In this case the procurement expert should be able to declare each filter (c.f. KO-criteria), so that the OAM contains only the (for the responsible commission) acceptable objects with their attributes.
The efficient handling of such kind of database can be ensured through OLAP techniques (or in offline case with a pivot-table).
The prescribing of the topical prices into the OAM of the authorities should be provided by the commercial partners.
If the choosing among the potential offers (objects) could be simplified to finding the cheapest object with the expected attributes, then it could be said: we have to manage an auction without being forced to include mathematical aid.
The right price-performance ratio is mentioned very often, if something is to buy. This is the root of the problems in the public procurement processes. For a definition of the price-performance ratio (especially in case of n-dimensional attribute-sets) was never given. Thousands of students (and their business environment) was not able to build any definition (although they had this topic as a problem to solve) for the right price-performance ratio. The most decisions in the private procurement are making instinctive. The search procedures for the key word 'price-performance ratio' deliver results as follow (1 & 2 & 3). These definitions can handle only the ratio between one attribute and the price. Have we more then one attribute, to evaluate objects, we can not find any calculations schemes for them. Instead of a correct definition we have only a sign for a kind of inverse proportion.

The MY-X FREE can support the following basic definition layers (and also their finetuning) concerning the optimizing of the price-performance ratio: (s. basic flowchart /SmartDraw-Trial/):

  • 1. channel = searching for antagonisms (0. level / choosing strategy):The evaluator can decide - having a given tender with more or less legitimate expectation as additional charges, weights, fix-points, direction of attributes), whether he want to accept a winner basing on direct antagonism (left half of the flowchart = 1. channel). Direct antagonism can be detected, if an object have an attribute series (=performance), which is not better than a series of an other object, but for more price.
1. channel / A. level: If only one object can be found as 'underestimated' (estimated price > fact price), then this is the winner. The further objects can be evaluated as balanced (estimation = fact) or 'overestimated' (estimated price < fact price) from point of view of the customer.
1. channel / B. level: If there are more objects having a price advantage (=[(fact-estimation)/fact]<0), then the winner is, who has the maximum of this advantage.
1. channel / BA. level:If there are more objects having the identical relative price advantage, then the winner is who has a minimum fact price.
1. channel / BB. level: Should be more from the cheapest fact price with an identical price advantage, then the evaluator have to choose a randomized object (=RND).
1. channel / C. level: If there is no antagonisms in the OAM to detect, but the evaluators want to be consequent and to make decision basing on inconsistent object proportions (having often corruption in their background - c.f. 'king maker' offers), then it is to prove, whether some attributes with monotone impacts (c.f. perturbation) are to find (c.f. STEP STD).
1. channel / CA. level: If there was no attributes with constant stairs to detect, then either the decision strategy can be changed (c.f. 2. channel) or it should be made a lot to determine the winner.
1. channel / CB. level: If there are more perturbations, then it is necessary to start a new calculation after deleting the attributes (they had impacts in the previous iteration). Man should go back always to the calculation level (1. channel 1. level) with the reduced OAM till either it becomes necessary to make a lot or to change strategy (to avoid invalid tenders).
  • 2. channel = minimizing risks (0. level / choosing strategy): The evaluator of a tender can reject each offer having a 'kingmaker' characteristics (nowhere better, but more expensive). The kingmaker offers can be ensured by anybody: for the own offer is always well-known, therefore other offers with less performance, but a higher price can let be submitted without any risk. Hereby it is necessary to detect, whether there are direct antagonisms in the OAM? If yes, and the evaluator want only work with an OAM not having such corrupt objects, then each kingmaker offer should be rejected step by step before the OAM will be analyzed finding a winner. (right half of the flowchart = 2. channel))
2. channel / A. level: Should be only remained one object after the iterative rejecting phase, the tender can be evaluated as invalid.
2. channel / B. level: If the iterations deliver finally an OAM without direct antagonisms (that is: pair-comparison detected nothing or an Y0-model - having the price among the attributes - can be solved as error free: demo), then it should be proved, whether there are still indirect antagonisms in the OAM.
2. channel / BA. level: If there are some indirect antagonisms (= inconsistencies, which can not be detected through pair-comparisons), then (c.f. 1. channel) we should be able to find a winner among the underestimated offers.
2. channel / BB. level: If there is no indirect antagonisms (namely an estimation can be calculated for each offer being the same as the appropriate fact price), then a STEP IX process should be run to detect, which thresholds can be identified for the minimum and for the maximum prices.
2. channel / BBA. level: If there is no threshold (= beyond that the model have not more an error free status) to calculate, then it should be analyzed whether the model have perturbations or not?
2. channel / BBB. level: If at least one threshold is given, then we have to follow the next instructions: At first we have to identify the object having the closest distance between fact price and its minimum threshold.
2. channel / BBBA. level: Should only be one such an object, then it is the winner.
2. channel / BBBB. level: If more such ratio can be detected, then we have to identify the object having the widest distance between fact price and its maximum threshold.
2. channel / BBBBA. level: Should only be one such an object, then it is the winner.
2. channel / BBBBB. level: Should be exist more objects with identical price advantages, then we have to make a lot (RND).
2. channel / BBAA. level: If there is no perturbation in an OAM providing an error free status, then we have to make a lot (RND).
2. channel / BBAB. level: If there are some perturbations, then we have to delete each attributes having already impacts and we have to go back to the starting point of the 2. channel.
  • As it can be clearly seen, the objectivity-oriented similarity analysis may finish in form of a tie-situation, inasmuch as the OAM contains less data than needed to declare a correct price-advantage (being calculated as always it will). The necessity of a lot can be reduced, if the amount of the offers will be maximized. So that: it should be always recommended to prefer a variation analysis.
  • It should be also evident that the avoiding of kingmaker offers is only a decision on ethic level under the given social and economical circumstances.
  • The similarity analysis ensure a kind of stability because nobody might be able to make an arbitrary influence to each offer in a tender. Therefore nobody can forecast the winner, even if he delivers corrupt offers.
  • To enforce a expected constellation of indirect antagonisms with a prescribed winner, it has the same chances as a hit in the lottery in case of 'n' objects and 'm' attributes.
  • In such an iterative complexity the automated similarity analysis can not be avoided from the process of the tender evaluations. The manual handling of this type of problems can only be justified through didactic reasons.

Further details:
Antagonisms: Corrupt offer or dumping price?: In case of an OAM, at first such kind of objects (offers) should be identified, of which performances are not better (than the values of other objects) but they have higher prices. These can be declared basically as corrupt (or kingmaker) offers pointing out potential winner. BUT: Contrary to this ethic-based interpretation, these objects can be also defined as offers with dumping prices. Might be declared a minimal price reducing competition but ensuring the achievements?
Cartel-suspicion: Nowadays we can find reports often about Cartel-suspicions. Can not we declare a cartel through a automated mathematical algorithm based on the real data of market processes instead of a judgment?
Strategies of comparison: Direct antagonisms (called often as competition) can be derived by two ways: first through pair-comparison, second through similarity analysis (vö. demo).
It should be emphasized, that a kingmaker offer can never be declared as winner. Such kind of offers make only possible to ensure chances for other offers. Inasmuch as direct antagonisms could not be found in an OAM the evaluator can declare always weights, which ensure a winner position for arbitrary offer. This matter (included into each public procurement process) is responsible for litigations (c.f. misappropriation, corruption, setting back business processes through seemingly correct litigations). In opposite to the real situations: The regulations of the public procurements should ensure the efficiency from two points of view: first the best price-performance ratio, second the legal stability.
It is worth rejecting the overestimated offers from the OAM (from ethical point of view). In such kind of OAM offers with indirect antagonisms can be detected also. These constellations can be only formed in case of relative huge number of objects and relative little amount of attributes (c.f. demo before). This constellation arises not often in a real public procurement process - except when the ideal database with each potential offer (generating by permission authorities) will be used to filter the objects for a tender. This would lead to a new e-commerce world with automated decision support solutions. Till then the evaluator should force the tenderer so that they work out more variation of offers.
The similarity analysis is able not only to detect the direct and indirect antagonism, but also the rule system (stairs) in the background of the proved markets - having seemingly an price-performance equilibrium for each offer.
An OAM without antagonisms may have perturbations: namely the similarity analysis is always searching for the potential equilibriums. To ensure an error free model-status, it is not necessary each attribute (or stair) to individualize. Monotone sequences can be arisen anywhere in the stair-system. Therefore a similarity analysis provides always an indirect proof: Are we able to make a decision or not? The information in the background of the monotone stairs can not be analyzed on the direct way, more with the STEP-IX or STEP STD modules.
In the STEP IX module each fact price is moved down and up stepwise, till the equilibrium can be kept. The STEP STD module works always with the perturbations in order to find direct or indirect antagonisms.
The analysis having the price among the X-attributes can not be called as analysis of the price-performance ratio. Namely the necessary scaling effect can not be found for detecting equilibrium between the price and the performance-attributes (c.f. Y0-models). In case of such an evaluation strategy the offer with price=0 can not be called as winner! It might have more points, but it will be never suspected as an offer with dumping price.
In case of already started procurement processes the COCS-module is able to simulate weighting systems.
Finally it should be emphasized that the similarity analysis can not replace the business calculations, or impact studies trying to provide exact benefits for each alternative offer. It is not the same question: which fodder has the best price-performance ratio OR which fodder might produce the best gross margin? BUT: both question can be solved through similarity analysis!
It should be also declared, that a similarity analysis can provide randomly the same solution (c.f. stairs) as a cost-price calculation.

Details I.
Details II.