A MODEL OF CONTRACTOR SELECTION USING ANP AND RISK ASSESSEMENT ON A PROCUREMENT SERVICE

Transcription

1 A MODEL OF CONTRACTOR SELECTION USING ANP AND RISK ASSESSEMENT ON A PROCUREMENT SERVICE Endi 1) and Bambang Syairudin Master s Program in Management of Technology, Institut Teknologi Sepuluh Nopember Jl. Cokroaminoto 12A, Surabaya, 60264, Indonesia 1) endividia.mmt@gmail.com ABSTRACTS Along with the many business entities of construction services in Indonesia, the government as the owner and the budget, in this case the unit procurement services can improve the quality of the evaluation of the service provider, which it is one of the important aspects of national development. The purpose of this study is to design a model of decisionmaking system to boost effectiveness and efficiency in the process of evaluation of the contractor. The method used in this study is Decision Making Trial and Evaluation (DEMATEL), Analytic Network Process (ANP), and the method of Risk assessement, some of these methods is an alternative method that can be used to measure the effect of intercriteria decision, correlation interdependence, as well as assessing risks of each contractor selection decision criteria. Respondents were used in this study is the committee in Tulungagung district procurement services unit consisting of five assessors with expertise in the procurement of construction services. Analysis of the data showed that the best alternative to have quality deals that not only take into account the offer price, but also consider some other aspects related to the condition and the object of the project. Of the three bidders, contractors B being the best alternative compared between the contractor A and contractors C based on the values obtained from the ANP and the level of risk weighting obtained from Risk assessement. Keyword: Selection, Decision Making Trial and Evaluation Laboratory, Analityc Network Process, Risk Assessement, Procurement. INTRODUCTION Selection of contractors to be one of the crucial process for the project owner to be able to complete the project in accordance with the request (Singh & Tiong, 2005). Selection of contractors that may cause a lot of reworking, and demolition will increase the probability of project delays, cost overruns can even lead to bankruptcy (Hatush & Skitmore, 1998). The Working Group Unit Service Procurement or can be called ULP is one of the institutions that play an important role in improving the quality, effectiveness, and efficiency of the procurement of government goods and services, therefore the decision-making process is a common thing and be the core activities of the working group ULP Tulungagung, selection of the contractor in the tender evaluation process is the decision-making process involving multiple criteria decision, whereby potential contractors are measured and assessed according to the criteria and subkrtieria decisions already in the set (Nieto-Morote & Ruz-Vila, 2012). In this paper, a decision support system for selecting candidates for the service provider or contractor uses Decision Making Trial and Evaluation (DEMATEL), Analytic Network Process (ANP), and the method of Risk assessement. Where the value of large-small A-1-1

2 influence among criteria or sub-criteria are calculated using DEMATEL, penentukan weight of each criteria and sub-criteria decision ANP calculated using the method and scale of the risk of each sub-criteria decision calculated using methods Risk assessement. METHODOLOGY Selection of contractors is a decision-making process that is identical to the decisionmaking model group (El-Sawalhi & Rustom, 2007) and also involve more than one criteria or sub-criteria decision, which some criteria or sub-criteria have a reciprocal relationship or correlation between the single criteria with the other, therefore DEMATEL and ANP methods used in this study in order to accommodate these concerns, while respondents who use ULP is the committee of five people who are experts in the field. 1. DEMATEL Method Selection of contractors is a decision-making process involving more than one criteria or sub-criteria (El-Sawalhi & Rustom, 2007), concluded that some of the criteria used to have a reciprocal relationship or correlation between the criteria of interdependence with each other. DEMATEL has been successfully applied in some situations such as marketing strategies, and evaluation ecolearning (Chiu et al., 2006). DEMATEL methods can confirm the interdependencies between variables or criteria and limit the relationship that describes the characteristics essential in a system and development trend. DEMATEL stages of processing methods as follows: (1) Determining the average matrix (matrix A); (2) Calculating the initial normalized direct-relation matrix (matrix D); (3) Calculating total correlation matrix (matrix T); (4) Trheshold assign value to obtain a matrix-relations impact map. 2. ANP Method Anality network process is the development of Analytic Hierarchy Process developed by Thomas L. Saaty (Saaty, 1994). Analytic network process is a method of measuring the compound criteria used to obtain the priority of individual assessment results using a scale or the absolute value of the actual measurement results are normalized. ANP method can accommodate the complexity of a decision-making model in which there is an interaction, interdependence, and feedback among the criteria used (Sevkli et al., 2012). There are three types of correlations between elements in decision-making model, the first is the correlation between elements in a cluster, the second is the correlation between the different elements of the cluster, and the third is a reciprocal relationship between cluster above or below, this approach devised a supermatriks to be adjusted again with eigenvector to generate relative priority weighting of each criterion (Cheng & Li, 2007). The decision making process on ANP method consists of four stages as follows: a. The design of the model; At this stage, the problem is translated in the form of a network setruktur decision, defining the elements of criteria and sub-criteria to be used, as well as the corresponding relationships between the criteria or sub-criteria. b. Pairwise comparisons; Elements of such criteria or sub-criteria and compared using a paired comparison method based on the criterion of control and the corresponding relationships that have been defined, the criteria and sub-criteria are compared based on their level of importance. Here is a grading scale which ranges from grades 1-9: c. Design supermatriks; Is a positive matrix, where each element in the matrix presented the relationship between the two elements in the network keptusan, the A-1-2

3 output of this stage is the weight or importance of each element in the network's decision. d. Synthesising the decision; At this stage of normalization of the supermatriks from the previous stage to gain importance weight of each element of the decision. 3. Risk Assessement Method Risk analysis aims to determine how often the risk of it happening and how big an impact the risks that affect the activities of the project, risk analysis include methods to determine priorities for the some of the risks identified, where the organization can improve project performance effectively by focusing on priority risks (Susilo & Kaho, 2009). In this study, qualitative risk analysis methods used to establish the probability and impact of risk on each sub-criteria decision. the scale of the impact-scale use of discrete values with a range of 1-3 as in the following table: Table 1. Probability Scale Numerical Rating Judgement Preference Remarks 1 Low Rare 2 Moderate Maybe 3 High Almost Certainly Reference: Walke, 2011 Table 2. Impact Scale Numerical Rating Judgement Preference Remarks 1 Low Increase < 5% of the project value 2 Moderate Increase < 6-25% of the project value 3 High Increase > 25% of the project value Reference: Walke, 2011 Based on the description of the methods above, the initial stage is determination of criteria and sub-criteria decision through a focus group discussion (FGD), of the criteria and subcriteria that have been defined then determined the relationship between criteria and subcriteria decision methods DEMATEL, after determining the weights of the criteria and subcriteria using the ANP, the last is to assess the risk of sub-criteria decision to use methods of risk assessement. The results obtained from the analysis of these three methods is used as a reference rank contractor, or as a reference for the selection of service providers best alternative among the other bidders. RESULTS AND DISCUSSION Based krietia and sub-criteria are already set, look for relationships between criteria and sub-criteria DEMATEL method, the goal is to determine the effect among criteria and sub-criteria decision and at the same time as the reference treatment comparison matrix that will be settled using ANP method at a later stage. A-1-3

4 Table 3. Total influence matrix Total influence matrix is the amount of influence each attribute to another attribute which if the value in the table influence matrix is smaller than the overall average influence matrix itself then that attribute has no effect, Table 3 below shows that the value of the red shows these attributes does not have any influence on other attributes, or in other words the value of the attribute is lower than the average value of the entire matrix itself. Last Stages on DEMATEL method is to calculate the value of dispatchers and receivers obtained from the processing matrix linkages total. The value of each criterion must be measured in terms of rows and columns. Value index of total power of influence given and received by a criteria or sub-criteria decision (D+R), and the value of the tendency of the degree of influence of each of the criteria or sub-criteria decision (D-R), if the value (D-R) is positive then the criteria or sub-criteria such decision tend to be more affecting, if the value (D-R) is negative, then the criteria or sub-criteria are likely to be more influenced by other criteria or sub-criteria. Briefly value of D + R indicates a causal relationship between the criteria of one another, so that the greater the value of D + R, the stronger the relationship between these criteria. Value (D+R) and (D-R) can be seen in table 4. Table 4. Value dispatcher dan receiver Subcriteria D R D+R D-R Subcriteria D R D+R D-R C1-1 2,520 2,485 5,005 0,034 C3-1 2,009 2,758 4,767-0,748 C1-2 2,440 2,393 4,833 0,047 C3-2 2,390 2,635 5,025-0,245 C1-3 1,027 1,535 2,562-0,508 C3-3 1,456 2,050 3,506-0,594 C1-4 2,363 2,226 4,589 0,136 C4-1 2,381 1,624 4,005 0,757 C2-1 2,442 1,954 4,396 0,488 C4-2 2,851 1,360 4,211 1,491 C2-2 2,698 1,658 4,356 1,041 C4-3 2,443 1,504 3,947 0,938 C2-3 2,264 3,058 5,322-0,795 C5-1 2,338 2,474 4,813-0,136 C2-4 1,685 2,701 4,386-1,016 C5-2 1,044 1,935 2,978-0,891 The results of table 4 can be converted into a map of impact-digraph, where these maps represent a description of the position of sub-criteria decision that will affect or be affected by other sub-criteria decision. A-1-4

5 Figure 1. Impact-digraph From figure 1 above shows that not all sub-criteria decision-making affecting the other sub-criteria, sub-criteria are positive affect subcriteria whose value is negative. Impactdigraph above as well as the basis for network model ANP. In Table 4 and figure 1 shows that the sub-criteria technical specifications, sub-criteria method of execution of the work, and the sub-criteria of project performance previously dominated the subcriteria other, the result can be used as an insert procurement committee, which the sub-criteria other highly influenced by the sub-criteria are, and will certainly affect the quality of the evaluation of election service providers. In determining criteria weights using ANP, required input from respondents about the influence of perception among criteria or sub-criteria decision and perceived level of importance of each criteria or sub-criteria decision against the contractor selection process, in which the result can be seen in table 5 and 6. The results of the questionnaire are taken of the respondents will then be calculated the average size, to be arranged in a pairwise comparison matrix (W ij ). Pairwise comparison matrix will be processed into weight criteria or subcriteria. Table 5. Results of the questionnaire interest rate decision criteria Respondents Criteria 1 Criteria Average Technical competence Organizational skil ,600 Technical competence Health and Safety Enviorenment ,600 Technical competence Internal ability 1 2 1/ ,700 Technical competence Bid price ,800 Organizational skil Health and Safety Enviorenment ,800 Organizational skil Internal ability ,200 Organizational skil Bid price ,400 Health and Safety Enviorenment Internal ability 1/2 1 1/2 1/3 1/3 0,533 Health and Safety Enviorenment Bid price ,000 Internal ability Bid price ,000 Table 6. Results of the questionnaire interest rate sub-criteria decision KT Subcriteria 1 Subcriteria 2 Respondents Avera ge Technical specification Construction method ,800 Technical specification Schedule of work ,200 Technical specification The type and capacity of equipment ,200 Construction method Schedule of work ,600 Construction method The type and capacity of equipment ,600 A-1-5

6 Schedule of work The type and capacity of equipment 1/3 1/5 1/5 1 1/5 0,387 project organization Project management 1 1/3 1/5 1 1/5 0,447 project organization Construction performance 1 1/5 1/5 1/5 1/3 0,387 KO project organization Quality control 1/3 1/3 1/4 1/4 1/2 0,333 Project management Construction performance 1/3 1/5 2 1/3 1/5 0,613 Project management Quality control 1/5 1/5 1/5 1 1/5 0,360 Construction performance Quality control ,000 Safety program Mitigasi pekerjaan berisiko tinggi 1/5 1/5 1/2 1 1/3 0,447 K3 Safety program Emergency response plan program 2 1 1/ ,700 Mitigasi pekerjaan berisiko tinggi Emergency response plan program ,600 Financial capability Number of projects completed ,400 KI Financial capability Type of project completed 1/3 1 1/3 1/3 1/5 0,440 Number of projects completed Type of project completed 1 1/3 1/2 1/3 1/2 0,433 HP Price of critical work Price of the work is not critical ,800 Super matrix from pairwise comparison matrix arrangement called Unweighted Supermatrix. Unweighted Supermatrix then normalized to obtain Weighted Supermatrix, Supermatrix Weighted normalized again to obtain Limiting Supermatrix. Weight of subcriteria is a representation of Limiting Supermatrix. The next stage respondents were asked to rate each contractor based on the subcriteria decision, the results of the assessment of contractors is multiplied by the weight of each sub-criteria, Table 7 below shows the sum of the weights with the results of assessment of contractors, whereby the three contractors as an alternative decision, the naming of the three contractors replaced using the term alphabet A, B, and C. Table 7. Results of the assessment contractor Subcriteria Weight Contractor A Contractor B Contractor C Value Total Value Total Value Total Technical specification 17% 4,200 0,707 5,000 0,842 3,400 0,573 Construction method 10% 3,400 0,346 3,800 0,387 5,000 0,509 Schedule of work 4% 3,800 0,161 5,400 0,229 5,000 0,212 The type and capacity of equipment 8% 3,400 0,280 5,400 0,444 4,200 0,345 project organization 3% 3,400 0,087 3,800 0,097 3,400 0,087 Project management 5% 3,800 0,185 5,800 0,282 4,200 0,204 Construction performance 9% 3,400 0,319 5,800 0,544 5,400 0,506 Quality control 6% 4,200 0,264 3,400 0,213 5,000 0,314 Safety program 5% 3,400 0,159 4,200 0,197 4,200 0,197 mitigation of high risk work 7% 5,800 0,379 4,200 0,275 3,800 0,249 Emergency Respnse Plan program 3% 3,800 0,111 4,200 0,122 3,800 0,111 Financial capability 5% 5,800 0,264 5,400 0,246 3,800 0,173 Number of projects completed 3% 4,200 0,145 3,800 0,131 5,400 0,186 Type of project completed 7% 3,400 0,242 4,600 0,327 4,600 0,327 Price of critical work 5% 3,800 0,202 4,200 0,223 3,800 0,202 Price of the work is not critical 3% 5,000 0,145 3,400 0,099 4,200 0,122 Total 100% 3,995 4,667 4,315 The analysis shows contractors B dominate other contractors on each sub-criteria decision, for example in sub-criteria technical specifications and subcriteria construction, kontraktro B gets top marks, the sub-criteria technical specifications, the contractor B obtain the value of 0.842, the next is A contractor with a value of 0.707, and the The latter is a contractor C has a value of In the sub-criteria of construction, contractor B scored higher than the contractor A and contractor C, contractor B has a value of 0,544, the next is the contractor C with a value of 0.506, and the latter is contracting A to obtain the value of A-1-6

7 The value of each contractor on each of the criteria is not always in the position of the same level, but the highest level is the contractor B. The last stage is the assessment of risk, each contractor has a different risk value to each sub-criteria decision, risk assessment is the product of the probability value obtained through questionnairesscale used are shown in Table 1, while the risk impact value Obtained through focus group discusion, scales used are shown in Table 2. The risk assessment is based on sub-criteria decision, Table 8 shows the results of a risk assessment of each contractor to each sub-criteria decision. Table 8. Results of the risk assessment Contractor A Contractor B Contractor C SUBCRITERIA Impact Probab ility Total Probab ility Total Probab ility Total Technical specification 3 1,600 4,800 1,000 3,000 1,600 4,800 Construction method 3 1,800 5,400 1,200 3,600 2,200 6,600 Schedule of work 2 1,400 2,800 1,000 2,000 2,200 4,400 The type and capacity of equipment 3 1,600 4,800 1,600 4,800 2,200 6,600 project organization 2 2,200 4,400 1,400 2,800 1,600 3,200 Project management 2 2,000 4,000 1,200 2,400 2,000 4,000 Construction performance 3 1,600 4,800 1,400 4,200 1,600 4,800 Quality control 3 1,400 4,200 1,400 4,200 1,400 4,200 Safety program 2 2,000 4,000 1,800 3,600 2,000 4,000 mitigation of high risk work 1 2,200 2,200 2,200 2,200 2,000 2,000 Emergency Respnse Plan program 1 1,800 1,800 1,200 1,200 1,800 1,800 Financial capability 2 1,600 3,200 1,000 2,000 1,800 3,600 Number of projects completed 1 1,200 1,200 1,400 1,400 1,800 1,800 Type of project completed 2 1,600 3,200 1,400 2,800 1,600 3,200 price of critical work 2 1,400 2,800 1,400 2,800 1,800 3,600 price of the work is not critical 1 2,200 2,200 1,600 1,600 1,600 1,600 Total 55,800 44,600 60,200 The results of the risk assessment is used as a divider of any value on the assessment of the contractor, and these results as well as a reference to rank each contractor. Value determination of the risk of contracting symbolized by CV, and RV votes symbolized by the contractor. The result can be seen in table 9. Table 9. Total value of contractors Subcrite ria Contractor A Contractor B Contractor C CV RV CV/RV CV RV CV/RV CV RV CV/RV C1-1 0,707 4,800 0,147 0,842 3,000 0,281 0,573 4,800 0,119 C1-2 0,346 5,400 0,064 0,387 3,600 0,107 0,509 6,600 0,077 C1-3 0,161 2,800 0,057 0,229 2,000 0,114 0,212 4,400 0,048 C1-4 0,280 4,800 0,058 0,444 4,800 0,093 0,345 6,600 0,052 C2-1 0,087 4,400 0,020 0,097 2,800 0,035 0,087 3,200 0,027 C2-2 0,185 4,000 0,046 0,282 2,400 0,117 0,204 4,000 0,051 C2-3 0,319 4,800 0,066 0,544 4,200 0,129 0,506 4,800 0,105 C2-4 0,264 4,200 0,063 0,213 4,200 0,051 0,314 4,200 0,075 C3-1 0,159 4,000 0,040 0,197 3,600 0,055 0,197 4,000 0,049 C3-2 0,379 2,200 0,172 0,275 2,200 0,125 0,249 2,000 0,124 C3-3 0,111 1,800 0,061 0,122 1,200 0,102 0,111 1,800 0,061 C4-1 0,264 3,200 0,083 0,246 2,000 0,123 0,173 3,600 0,048 C4-2 0,145 1,200 0,121 0,131 1,400 0,094 0,186 1,800 0,103 C4-3 0,242 3,200 0,076 0,327 2,800 0,117 0,327 3,200 0,102 C5-1 0,202 2,800 0,072 0,223 2,800 0,080 0,202 3,600 0,056 C5-2 0,145 2,200 0,066 0,099 1,600 0,062 0,122 1,600 0,076 Total 1,213 1,684 1,176 A-1-7

8 Where the value of CV obtained from the analysis of the contractor assessment and RV values obtained from the analysis of risk assessement, the division between the contractor value and the risk value is the final assessment of the contractor. CONCLUSIONS AND RECOMMENDATIONS From the overall results of this research process, indicating that the sub-criterion technical specifications are sub-criteria that gets the highest weight, based on the analysis results obtained weighting of 17%, the second priority is the method of execution of the work, the weight is obtained 10%, the third priority is the construction and then followed by several other sub-criteria. Based on the analysis, contractor B gets the highest score, the results obtained was Based on the draft model of contractor selection criteria based compound method, the contractor obtained winners have quality deals that not only take into account the offer price, but also consider the technical capabilities, organizational capabilities and several other aspects of the conditions and objects in the project. Therefore, the selection of the contractor in this case study is more appropriate to use merit point method than knockout method, because the method is closer to the facts. The impact of risk in this study was limited to the side of the owner, and the expectations on future research is in order to measure the impact of risk on the contractor side. The second recommendation is about the orientation of the risk impact, not only measured in terms of cost, but also in terms of quality, scope, and time, so that the risk assessment is to provide an overview in detail. REFERENCES Cheng, E. W., & Li, H Application of ANP in Process Models: An Example of Strategic Partnering. Building and Enviorenment, 42, Chao-Min, C., Meng-Hsiang, H., & Eric, T.G, W Understanding Knowledge Sharing in Virtual Communities: An Integration of Social Capital Cognitive Theories, 42, El-Sawalhi, N. E., & Rustom, R Contractor Prequalification Model: State-Of-The-Art. International Journal of Project Management, 25, Hatush, Z., & Skitmore, M Contractor Selection Using Multicriteria Utility Theory: An Additive Model. Building and Enviorenment, 33 (2-3), Nieto-Morote, A., & Ruz-Vila, F A Fuzzy Multi-Criteria Decision-Making Model For Construction Contractor Prequalification. Automation in Construction, 25, Saaty, T. L How to make a decision: The Analytic Network Process. Decision Analysis, 24, Sevkli, M., Oztekin, A., Uysal, O., Torlak, G., Turkyilmaz, A., & Delen, D Development Of A Fuzzy ANP Based SWOT Analysis For The Airline Industry In Turkey. Expert System with Applications, 39 (1), Singh, D., & Tiong, R. L A Fuzzy Decision Framework For Contractor Selection. Journal of Construction Engineering and Management, 62, Susilo, L. J., & Kaho, V. R Manajemen Risiko Berbasis ISO PPM Manajemen. Walke, R. C An Approach To Risk Quantification In Construction Project Using EMV Analysis. International Journal of Engineering Science and Technology, 3. A-1-8