Using multi-attribute decision-making approaches in the selection of a hospital management system

Abstract

The most appropriate organizational software is always a real challenge for managers, especially, the IT directors. The illustration of the term “enterprise software selection”, is to purchase, create, or order a software that; first, is best adapted to require of the organization; and second, has suitable price and technical support. Specifying selection criteria and ranking them, is the primary prerequisite for this action. This article provides a method to evaluate, rank, and compare the available enterprise software for choosing the apt one. The prior mentioned method is constituted of three-stage processes. First, the method identifies the organizational requires and assesses them. Second, it selects the best method throughout three possibilities; indoor-production, buying software, and ordering special software for the native use. Third, the method evaluates, compares and ranks the alternative software. The third process uses different methods of multi attribute decision making (MADM), and compares the consequent results. Based on different characteristics of the problem; several methods had been tested, namely, Analytic Hierarchy Process (AHP), Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), Elimination and Choice Expressing Reality (ELECTURE), and easy weight method. After all, we propose the most practical method for same problems.

Keywords

Software selection criteria multi-attribute decision-making (MADM)buying model buy vs. make organizational buying behavior (OBB)hospital management system

1. Introduction

Organizational buying is decision making process during which organizations make the need for goods and services and then start evaluating, assessing and selecting from sellers and suppliers. One of the most important challenges of the organization’s manager of information technology (henceforth, IT) is selecting appropriate software considering the organization’s conditions among several items and various software companies [23]. The current attempt will investigate the steps a company must go through to buy the software and explore ways of selecting the best software from several options. The significance of this project will be more highlighted when one refers to statistics obtained at 2008 that more than 25% of IT projects failed and more than 44% faced challenges [13]. To date, various models have been proposed for evaluating the purchase in contrast to making software in the organization. Some of these models deal with comparing the purchase in contrast to the in-house development, from which one can point to Cortellessa’s study as one which offers a mainstream in which decisions about buying or producing software are made based on tangible or non-tangible costs [6]. Based on his proposed model, Buchowicz will decide about buying or making the software by the company [2]. Platts compares in-house development with outsourcing on the basis of Multi Attribute Decision Making (MADM) [26].

MADM method applied in material selection or activity evaluation [8, 29, 19, 15, 17]. Another paper presents a material selection approach for selecting absorbent layer material for thin-film solar cells (TFSCs) using MADM approach [8]. Kumar and Agrawal propose a methodology by which selection of electroplating product plant can be made easy by MADM method. His procedure will help the user to select the system most suited for his operational needs. Moreover, his paper discusses how the electroplating suppliers, designers and maintenance personnel will also be benefited [19]. Venkateswara and Baral’s paper describes a methodology for evaluation, comparison, ranking and optimum selection of a feed stock for anaerobic digestion by MADM method [32]. Huang et al. present a new multi-criteria decision making (MCDM) model and uncertainty analysis method for the environmentally conscious materials selection problem. His model is based on engineering analysis and life cycle simulation and the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method is employed, and uncertainty analyses are performed for model flexibility. His model is applied to material selection for PC housing [15].

Bernroider’s article seeks to enhance acceptance of the profile distance method (PDM) in decision support systems. The PDM is a multiple attributive-based decision making as well as a multiple method approach to support complex decision making and uses a heuristic to avoid computationally complex global optimization [3]. Water and Peet introduce decision support system (DSS) which selects the software for factories based on the hierarchical decision-making approach [34]. Hwanga et al. introduce web-based DSS which suggests one of these two solutions: in-house development or buying the software, based on MADM [14]. However, organizations tended to outsource their organizational software due to the lower costs and for using others’ best practices and the models proposed for comparing production with buying software were not useful in these organizations. For instance, Juha and Pentti [18] talk about the ways organizations can manage risks in their purchases. Rainer et al. investigate some cases of purchasing software and finally suggests a process for reducing the risks of buying software [27]. Wybo et al. suggest a method in which the best item can be selected based on the search [35]. Güngör et al. introduce DSS in which qualitative and quantitative indicators are categorized [11]. However, web-based solutions have some potential issues, which need to be careful considered about them [20, 37]. Also, some models have been proposed for buying special software. From these models, Ariyachandra and Watson’s study can be named as one which identifies the software selection indicators of data warehouse [1] or to Verville and Halingten’s writings which specify the steps and indicators of selecting Enterprise Resource Planning (ERP) 1 software [33]. Nevertheless, to date, a model which can help the manager in buying any kind of organizational software has not been proposed and this is an issue which is under investigation in the present article. Finally, it will be tested and selected as a case study of the university system for one of the Iran’s private organizations. Accordingly, in the second part, a model for selecting organizational software will be presented. Then we will investigate the comparison indicators and select the organizational software. At the end of this stage, a classification of the indicators for selecting organizational software will be presented. Finally, in the last stage, a comparison will be made among various items using MADM and the selection priorities will be identified.

2. Literature review

2.1 Organizational buyer behavior

Organizational buying is a decision making process during which organizations feel the need for goods and services and then start investigating, assessing and selecting from various existing sellers and suppliers of goods and services. Customer’s behavior and organizational buying behavior models which are created in the last four decades show intuitive methods of decision making and customers’ behavior. The main reason of developing organizational buying behavior models is to present a theory and facilitate learning of what is presented [21]. Models are used as a useful tool to assist management decision-making in marketing operations; but due to the dynamicity of markets, buying behavior models should be constantly reviewed and updated. Factors that influence organizational buying are environmental factors, marketing incentives, in-house factors, risk tolerance and group and individual factors [21]. Several models have been proposed for organizational strategic buying. Models of Ghingold and Wilson [12], Webster and Wind [36], Seyedaghaee et al. [24, 25], Robinson et al. [28] and Burger and Cann [5] are among the most famous ones.

2.2 Organizational software selecting indicator

Indicators and criteria can be considered as a tool for defining, monitoring and evaluating the quality or selection [7, 27, 20]. The concept and notion of criteria shows the principals or main conditions for the stability of decision. Each criterion has several qualitative and quantitative indicators for monitoring and measuring which are regularly and constantly measured to determine and specify the effects of decision. An indicator is a variable which is used for evaluating conditions and sensitivities, comparing choices, evaluating conditions and preferences, and predicting conditions and future trends [7]. Table 1 shows the indicators’ valuation based on two-paired comparison.

Table 1
Indicators’ valuation based on two-paired comparison

Preferred value	Comparison of $i$ to $j$	Explanation
1	Approximately equal	Option or indicator $i$ and $j$ are of equal importance or none is superior.
3	Relatively more	Option or indicator $i$ is slightly more important than $j$ .
5	More	Option or indicator $i$ is more important than $j$ .
7	Much more	Option or indicator $i$ is much more important than $j$ .
9	Much much more	Option or indicator $i$ is absolutely more important than $j$ and is not comparable to $j$ .
2, 4, 6, 8	Intermediate value	The intermediate values between the preferred values.

2.3 Multi-attribute decision making (MADM)

MADM system is a well-established expert decision system when there exist conflicting criteria to evaluate a product/system among the available options. MADM is a decision making tool that enables the rigorous selection of the most preferred choice in a context where several criteria apply simultaneously. In a rational decision making environment, the most preferred choice is generally bounded by the management objectives, and the constraints that limit the choices and the achievement of the objectives [9]. MADM methods include a series of techniques like weighting or the convergence analysis which allow to weight and rate a range of criteria related to a topic to be further ranked by experts and interest groups. MADM methods are classified into two groups of compensatory and non-compensatory [38]. In non-compensatory methods, if an item does not meet the indicator, it will be excluded, while in the compensatory methods, all indicators are given a point and finally the indicator with the highest point will be selected. Among non-compensatory MADM methods, lexicographic, dominance, maxi-max, maxi-min, sufficiently good and conjunctive-sufficiently good, disjunctive-satisfying methods and conjunctive-satisfying methods are more fames and Among compensatory MADM methods, the most important are AHP,1

¹
Analytic Hierarchy Process.

TOPSIS, ELECTRE, Simple Additive Weight (SAW),2

Simple additive weighting.

VlseKriterijuska Optimizacija I Komoromisno Resenje (VIKOR), AV,3

Value analysis.

etc. The main steps of multi-criteria decision making are listed as follows the following:

(a)

Establishing system evaluation criteria that relate system capabilities to goals.

(b)

Developing alternative systems for attaining the goals (generating alternatives).

(c)

Evaluating alternatives in terms of criteria (the values of the criterion functions).

(d)

Applying a normative multi-criteria analysis method.

(e)

Accepting one alternative as optimal (preferred).

(f)

If the final solution is not accepted, gather new information and go into the next iteration of multi-criteria optimization.

One of the main problems of various compensatory MADM is that they will yield different answers for the same problem. Scientifics found that in at least 40 percent of cases, the result of each approach differed from the others. This nonoccurrence might have different reasons, like difference in weighting criteria, difference in achieving the best option in various approaches, the fact that some approaches use goal ranking instead of applying their weights and ultimately the fact that in some approaches extra parameters are defined which affect the results. Thus, different approaches are proposed for different problems [16].

Analytic Hierarchy Process (AHP) method is one of the most famous multi-purpose decision-making techniques which were first invented by Saaty [30]. This method lies on paired comparisons. The procedure for using the AHP can be summarized as:

(1)

Model the problem as a hierarchy containing the decision goal, the alternatives for reaching it, and the criteria for evaluating the alternatives.

(2)

Establish priorities among the elements of the hierarchy by making a series of judgments based on pairwise comparisons of the elements. For example, when comparing potential real-estate purchases, the investors might say they prefer location over price and price over timing.

(3)

Synthesize these judgments to yield a set of overall priorities for the hierarchy. This would combine the investors’ judgments about location, price and timing for properties A, B, C, and D into overall priorities for each property.

(4)

Check the consistency of the judgments.

(5)

Come to a final decision based on the results of this process.

TOPSIS is another MADM method which was originally developed by Hwang and Yoon in 1981 [16]. TOPSIS is based on the concept that the chosen alternative should have the shortest geometric distance from the positive ideal solution and the longest geometric distance from the negative ideal solution. The TOPSIS procedure consists of the following steps:

(1)

Calculate the normalized decision matrix.

$\displaystyle n_{ij}=\frac{\textit{ajj}}{\sqrt{\sum\limits_{i=1}^{m}\textit{% ajj}^{2}}}$ (1)

(2)

Calculate the weighted normalized decision matrix.

$\displaystyle p_{ij}=\frac{a_{jj}}{\sqrt{\sum\limits_{i=1}^{m}a_{jj}}}$ (2)

(3)

Determine the ideal and negative-ideal solution.

$\displaystyle d_{j}=1-E_{j}$ (3)

(4)

Calculate the separation measures, using the n dimensional Euclidean distance.

$\displaystyle d_{j}^{+}=\sqrt{\sum_{j=1}^{m}(v_{ij}-v_{j}^{+})}^{2}\quad d_{j}% ^{-}=\sqrt{\sum_{j=1}^{m}(v_{ij}-v_{j}^{-})}^{2}$ (4)

(5)

Calculate the relative closeness to the ideal solution

$\displaystyle\textit{CLi}=\frac{dj^{-}}{dj^{-}-dj^{+}}$ (5)

(6)

Rank the preference order [16].

ELECTRE is a family of MADM methods that originated in Europe in the mid-1960s. The nature of the recommendation depends on the problem being addressed: choosing, ranking or sorting. Criteria in ELECTRE methods have two distinct sets of parameters: the importance coefficients and the veto thresholds. The procedure of ELECTRE is:

Calculate the normalized decision matrix

Calculate the weighted normalized decision matrix

Determine the concordance and discordance set

$\displaystyle C(p,q)=\{j|\textit{vpj}\geqslant\textit{vqj}\}$ (6) $\displaystyle D(p,q)=\{j|\textit{vpj}<\textit{vqj}\}$

Calculate the concordance matrix and the discordance matrix

Determine the concordance dominance matrix

Determine the discordance dominance matrix

Determine the aggregate dominance matrix

Eliminate the less favorable alternative [38].

The Simple-additive-weighting method (SAW) is the simplest MADM method for evaluating a number of alternatives in terms of a number of decision criteria. Suppose that $w_{j}$ denotes the relative weight of importance of the criterion $C_{j}$ and $a_{ij}$ is the performance value of alternative $A_{i}$ when it is evaluated in terms of criterion $r_{j}$ . Then, the total importance of alternative $A_{i}$ , denoted as $A_{i}^{*}$ , is defined as follows:

$\displaystyle A^{*}=\left\{{A_{t}|\max\limits_{i}\frac{\sum\limits_{j}W_{j}% \cdot r_{ij}}{\sum\limits_{j}W_{j}}}\right\}$ (7)

For the maximization case, the best alternative is the one that yields the maximum total performance value [31].

3. The proposed approach

In selecting the software, first organization’s needs should be identified. Then, based on the need analysis, required modules can be determined. Then, the main companies in this field should be found and the need proposals (Request For Proposal (RFP)) should be sent to them to receive the initial price. If none of the companies could fulfill the needs or if their costs were higher than production costs, the solution of producing from outset (whether in-house or out-house) will be selected. Since organizations have similar software requirements, in most cases the buying option will be selected. Finally, one option should be chosen from the proposed solutions and different software. To choose among several options, a comparison should be notably made to find the best item; this comparison needs some criteria based on needs to be able to select the best item. A software buying model for an organization which covers all these steps is presented in Fig. 1.

Figure 1.

The model of buying organizational software.

According to this model, the software should be selected from a number of options based on software’s general indicators and the modules specified in the needs analysis stage. Accordingly, a number of indicators should be determined and based on some methods, the options should be compared.

3.1 Evaluating indicators selection

Software indicators are diverse and great. They are discussed in several books and articles. This article has evaluated 41 various articles and books to investigate indicators related to software and has extracted selection indicators from them. First, these indicators should be general and applicable for each type of organizational software; second, they should be understandable and measurable. For better investigation, the indicators were put in three categories. The first was general software selection indicators, the second was manufacture and seller’s indicators and the third was factors and standards of software quality.

Table 2
Software selection indicators

General software selection indicators	Process-oriented, supporting multiple languages, modularity, flexibility, multimedia content, ease of operation, changeability, easy use, multiple version, understandability, documentation of all affairs, meeting future needs, having software quality standards, user-friendly environment, meeting all current needs, using the best practice methods, appropriate and step-by-step help, compatibility with software, operating systems and browsers, security, no need for other software, easy manipulation, user interface standards, the amount of resources involved in executing a command, integration of the whole system, up to date, having the program text, open source, easy installation
Company and seller’s indicators	History of software, the annual growth rate, support, education, market share, customer’s significance, financial strength, number of employees, successful performance in similar organizations, department of research and development, company’s background and reputation, having a set of organizational software for future needs, customer satisfaction, level of costs, installation and deployment time, localization, physical vicinity
Factors and standards of software quality assessment	Architecture, cost-effectiveness, efficiency, functionality, facility, confidence, support, maintenance, cooperation, transportation, flexibility, integration, universality, maturation, operationally, testability, global, responsibility, self-sufficiency, accuracy, adaptability, correctness, reusability

As indicated in Table 2, there are a very large number of indicators and their ranking and investigation is a time-consuming and difficult task. Additionally, the selection indicators have different values and all do not have similar values in the selection procedure; so, if possible, indicators of low importance should be excluded. Moreover, some indicators cover other indicators and consequently most indicators will be combined or excluded. Hence, the final indicators were selected among these indicators. To select the final indicators, we went through the following steps:

Removal of indicators of low importance: those indicators that are not in most models are of low significance.

Removal of similar indicators: indicators which have high commonalities are one in fact and follow the same objective. So, repeating them is a mistake which confuses the addressee.

Removal of specific indicators: those indicators which are not general and gain importance in one or two software have no special position in the set of indicators.

Categorizing indicators: indicators have not equal importance and their weights are different in different organizations based on organization’s characteristics. For this reason and for the ease of comparison between options, the indicators were categorized. Of course, it should be noted that the weight of each category is different in various organizations.

Figure 2.

Final indicators of software selection.

Figure 2 shows this categorization. As illustrated in the figure, the indicators are categorized into four main indicators of technical indicators (A), user and functional indicators (B), management indicators (C) and cost indicators (D) and 15 sub-indicators which are sub-branches of main indicators.

4. Comparing the items and selection (Implementation)

After determining the final criteria for evaluation, the approbatory of such criteria should be investigated. Therefore, these criteria were evaluated in a case study (University management software). So that one of the Iran’s private organizations has a university-like structure and accepts many students each year, needs comprehensive university system for automating, maintaining discipline and accelerating the operations related to students. To this purpose, models and indicators presented in this article are used for selecting the university system in this organization. According to the model of Fig. 1, needs of this organization have been determined in the needs analysis stage and in the data collection stage, 7 active companies have been found in this field; then, the characteristics and costs have been inquired in a proposal. From these 7 companies, some will not fulfill organization’s initial needs and have to be excluded. As illustrated in the model, this is done in the filtering stage. To do this, first the system’s required modules should be determined. These modules are categorized into two groups of functional and technical indicators. Functional indicators are sections and modules which are expected to support the system and technical indicators are facilities and characteristics of the program which the software system should possess to be able to be run in the organization. These indicators are prerequisites for the entrance of software to the comparison stage; in the sense that if the software does not support these indicators, it will automatically be removed from the selecting options. In Fig. 1, the functional indicators of university system, which is required in the mentioned organization, have been specified.

Figure 3.

User and functional indicators of software selection in an organization.

In most cases, technical characteristics of software are attended besides functional modules in organizations; characteristics which fulfill most technical needs and make the software dynamic and flexible. In the mentioned organization, the following technical characteristics have been needed which are specified in Fig. 4.

Figure 4.

Technical indicators of software selection in an organization.

4.1 Implementation method

Different selection methods have been proposed based on indicators. From those one can point to fuzzy methods, artificial intelligence methods, multi-criteria selection methods, etc. Here, we used some clustering methods to identify needed features [4] Multi-attribute decision-making methods seem to be the best; since firstly, we can simply consider all indicators and determine individual weights for each of them. Secondly, these methods are simple, available and common for problems and thirdly it proposes the selection priority instead of selecting an option. It is noteworthy that using MADM methods in this problem is for the purpose of disciplining the selection and helping the organization’s manager in selecting the organizational software; selection priority is proposed in different conditions and in the final stage the manager has the right to select.

4.2 Implementation stage

4.2.1 First stage (filtering)

As said before, in the primary search, various items are proposed and some of the items do not cover main and primary needs of the organization. Thus, in the initial stage software that do not fulfill organization’s primary needs are excluded. Based on this (excluding some options), non-compensatory (irreversible) MADM methods are used. Among non-compensatory MADM methods, lexicographic, dominance, maxi-max and maxi-min are not applied since these methods determine the priority among indicators; while in this problem there is no special priority or ordering of indicators. Also, the sufficiently good and disjunctive-satisfying methods are not applicable. From the two methods of conjunctive-sufficiently good and conjunctive-satisfying methods, the conjunctive-sufficiently good was selected. Since in the conjunctive-satisfying method the options should cover all indicators but in the conjunctive-sufficiently good method a standard (sufficiently good) is determined for each criterion which is called the standard level. The accepted option is one which does not deviate from the standard level in all indicators [16]. Thus, we first determined the extent to which software options covered the evaluation criteria. Additionally, we determined sufficiently good. If the software’s weight was lower than sufficiently good in at least one of the indicators, it was excluded. The indicator’s weight is one of these values: it possesses i.e. full coverage, it does not i.e. no coverage and to some extent which means relative but not full coverage. Since these weights were of string type, they were transformed to numerical equivalent (possesses $=$ 1, not possesses $=$ 0 and to some extent $=$ 0.5). Since none of the software option could completely fulfill the organization’s needs, needs were restricted to system’s main required modules and the above task was repeated. The result of this stage is illustrated in Table 3.

Table 3
Filtering items based on sufficiently good method

	“S1”	“S2”	“S3”	“S4”	“S5”	“S6”	“S7”	Minimum
								acceptable
Student evaluation	1	1	0.5	LTM ${}^{*}$	0.5	0.5	1	0.5
Pre-registration, registration,	1	1	1	LTM	1	1	1	1
weekly program, scoring
and basic operations
associated with education
Attendance	0.5	0	1	1	0	0.5	0	0
Admission	0.5	1	1	LTM	1	0.5	1	0.5
Accommodation	0.5	1	0	0.5	1	0.5	0	0
University acceptance test	0	1	0	0	0	0	0	0
Student financial services	1	1	1	1	1	1	1	1
Graduation	0.5	1	1	LTM	1	0.5	1	0.5
Teacher evaluation	1	1	1	LTM	1	1	1	1
Teachers tuition	0	1	0.5	LTM	0.5	1	LTM	0.5
Letter generator	1	0.5	0	0	0	1	0	0
Report generator	1	1	1	LTM	LTM	1	LTM	0.5
Work flow designer	0	1	0	0	0	1	0	0
Export to Excel, Access, xml	1	1	0.5	LTM	LTM	1	LTM	0.5
or other database
Form generator	0	1	0	0	0	1	0	0
Web based	0.5	1	1	LTM	1	1	1	0.5
The overall result	9.5	14.5	9.5	Elimination	Elimination	12.5	Elimination

${}^{*}$ Less than minimum (LTM).

Finally, 4 software options were extracted. From these 4 options, software “S.6” was excluded because of legal problems. From the remaining options, software which covered the needs, i.e. “S.1”, “S.2” and “S.3”, were selected and were entered into the comparison stage.

4.2.2 Second stage (comparison and selection priorities)

Now, the main problem is proposed that which option is more appropriate for the organization from these options that all cover the organization’s basic needs. In many cases the software selection has intuitive progress. In some cases, the final option is determined as we specified the indicators for different software and there is no need for comparison. In some cases, if some options satisfy our initial needs, we will choose one with lower cost. In some other cases, the one which is faster and easier is selected according to the manager’s discretion and if it fell short, we will go to the next option. The method followed in this attempt is compensatory MADM. Since first of all, this method excludes no option and considers all. Secondly, there will be an ordering of options and thirdly it is a disciplined method for making a comparison between the options.

In compensatory MADM, the specified indicators will be given a score which determines the importance of the indicator considering the organization’s need. Additionally, a number will be given to each option (“S.1”, “S.2” and “S.3” software in this case) which shows the weight of the option in the respective indicator. Finally, the scores of various options will be compared and the order of software selection will be specified. There are several compensatory MADM each of which has its appropriate place with a slight difference.

The validity of an approach in a problem will be determined based on its ability to reflect the decision maker’s wills. For example, an approach is appropriate for optimistic decisions while the other is good for pessimistic ones and the result of applying them in a similar problem might be different; however, it would not be the reason for the inefficiency of any of them. TOPSIS is suitable when the nature of the problem accepts a very low risk; since the suitable option for this method is one which has the highest distance from the negative limit and lowest from the positive limit. In general, this method is used for testing the accuracy of the results of other methods. The weighted average is the oldest and simplest method. AHP is more appropriate for hierarchical problems or those in which criteria (indicators) are categorized in different classes and does not have the problems of the weighted average. ELECTRE is one of the best techniques of MADM which generally does not end in ranking the options solely; rather it might specify the best options and it might be called the top ranking approach. The AHP method is applied more in America and the ELECTRE method has been more seen in European studies [38]. Due to its hierarchical nature, special rankings of indicators and the closeness of Iranian thoughts and organizational models to Americans, AHP method was used as the main approach and ELECTRE, TOPSIS and weighted average as the subsidiary ones.

For weighting the options two-paired comparison, weighting based on the highest score and Entropy can be used. Among these approaches, Entropy cannot be used for our study, due to the nature of the problem. Two-paired comparison makes the comparison between the approaches easier and it can be easily changed to ranks; thus, this method was used for weighting. For each 15 sub-indicators and 4 main indicators two-paired comparisons were carried out. For one of the indicators, the two-paired comparison table and the preferred value table were put according to Tables 4 and 5.

Table 4
Two-paired comparison of take log, track events and improve matters (A1)

A1	“S1”	“S2”	“S3”
“S1”	Approximately equal	–	–
“S2”	Much more	Approximately equal	More
“S3”	Relatively more	–	Approximately equal

Table 5

Justified value of log indicator

A1	“S1”	“S2”	“S3”
“S1”	1	1/7	1/3
“S2”	7	1	5
“S3”	3	1/5	1

The sum of the numbers of each column of the paired comparison matrix is calculated and each element of the column is divided to the sum of its column. The resulted matrix is called “normalized comparison matrix” as shown in Table 6.

Table 6

Normalized comparison matrix

“S1”	“S2”	“S3”	Normalized
0.09	0.11	0.05	0.08
0.64	0.75	0.79	0.72
0.27	0.15	0.16	0.19

The average of the numbers of each row of the normalized matrix is calculated.

This average provides the relative weight of decision elements with the rows of matrix.

Table 7

The relative weight of options and sub-indicators

	S1	S2	S3	Sub-indicators		S1	S2	S3	Sub-indicators
				weight					weight
A1	0.08	0.72	0.19	0.14	C1	0.09	0.3	0.61	0.23
A2	0.06	0.65	0.29	0.24	C2	0.06	0.7	0.24	0.08
A3	0.07	0.64	0.28	0.08	C3	0.63	0.26	0.11	0.12
A4	0.18	0.75	0.07	0.54	C4	0.07	0.62	0.31	0.57
B1	0.19	0.72	0.08	0.21	D1	0.32	0.05	0.63	0.71
B2	0.1	0.69	0.21	0.58	D2	0.22	0.06	0.72	0.14
B3	0.09	0.61	0.3	0.13	D3	0.65	0.06	0.29	0.14
B4	0.19	0.72	0.08	0.07

Consequently, we can obtain the relative weight of options in other indicators. Also, the relative weight of sub-indicators of main indicators (for instance the relative weight of sub-indicators of user and functional indicators) will be obtained (See Table 7).

Then, the weight of each option should be multiplied by its indicator to find its weight with respect to the total indicators. Then, the sub-indicators’ weights are added to each other to obtain the weight of each option in the main indicator as shown in Table 8.

Table 8

Integrating relative weights of two-paired weighting

	S1	S2	S3		S1	S2	S3
A1	0.01	0.10	0.03	C1	0.02	0.07	0.14
A2	0.01	0.16	0.07	C2	0.00	0.05	0.02
A3	0.01	0.05	0.02	C3	0.08	0.03	0.01
A4	0.10	0.41	0.04	C4	0.04	0.35	0.18
A	0.13	0.71	0.16	C	0.14	0.51	0.35
B1	0.04	0.15	0.02	D1	0.23	0.04	0.45
B2	0.06	0.40	0.12	D2	0.03	0.01	0.10
B3	0.01	0.08	0.04	D3	0.09	0.01	0.04
B4	0.01	0.05	0.01	D	0.35	0.05	0.59
B	0.12	0.69	0.18

Then, based on the two-paired comparison of main indicators, the weight of each main indicator will be obtained. It should be noted that the weights of main indicators differ in different organizations.

Followed, weights of main indicators in organizations with average financial constraints are specified (See Table 9).

Table 9

The weight of first-level indicators in organizations of average cost

Technical indicators	0.07
User and functional indicators	0.15
Management indicators	0.39
Cost indicators	0.39

Finally, the weight of each main indicator is multiplied by each option to obtain the option’s final weight as can be seen in Table 10.

Table 10

Options’ final weight

Final weight	“S1”	“S2”	“S3”
A	0.01	0.05	0.01
B	0.02	0.11	0.03
C	0.05	0.20	0.14
D	0.14	0.02	0.23
Priority	0.22	0.37	0.41

To test the accuracy of the study’s results, TOPSIS was used. Since it is possible for the problem to be used in European organizations and to compare different methods SAW and ELECTRE were used as sub-methods. The results of these methods are provided in Table 11.

Table 11

Options’ final weight based on MADM

MADM	“S1”	“S2”	“S3”
AHP	0.22	0.37	0.41
Topsis	0.41	0.50	0.69
SWA	0.46	0.62	0.63
Average	0.36	0.50	0.57

Also these operations were calculated based on ELECTRE and results are specified in Table 12.

Table 12

Option’s ranking based on ELECTRE

	“S1”	“S2”	“S3”
“S1”	–	0	0
“S2”	0	–	0
“S3”	1	1	–

Removing low-gravity options, “S.3” software will be selected as the first option. As indicated in the above tables, based on the four methods of analytic hierarchy process, TOPSIS, ELECTRE and simple additive weighting and consequently based on the average of these methods, “S.3” software was selected as the first option.

As mentioned above, the weight of main indicators has a high influence on the selection of the final option; moreover, it was mentioned that this weight is different in different organizations. Thus, organizations are categorized into three groups of organizations with average financial constraints, organizations with high financial constraints and low risk and organizations with no financial constraints. The most influential indicator (criteria) in selecting different companies is financial constraints. Though, management indicator is also influential. Using analytic hierarchy process, the results were once calculated for organizations with low budget and low risk-taking power and once for organizations with no financial constraints:

Organizations with average financial constraints: in these organizations which include most Iranian organizations, management indicators and price are more important than technical and user characteristics.

Organizations with high financial constraints and low risk-taking power: organizations have special constraints in crisis; in addition, in some organizations due to budget problems or inappropriate position of IT, there are several financial constraints. In these organizations, proper price is the most important factor and has a high value in selection.

Table 13

The main indicators’ weight in organizations with high financial constraints

Technical indicators	0.06
User and functional indicators	0.15
Management indicators	0.28
Cost indicators	0.52

Table 14

Selecting the final option in organizations with high financial constraints

	“S1”	“S2”	“S3”
Final weight	0.25	0.31	0.44

Table 15

The main indicators’ weight in organizations with low financial constraints

Technical indicators	0.08
User and functional indicators	0.26
Management indicators	0.50
Cost indicators	0.16

Table 16

Selecting the final option in organization with low financial constraints

	“S1”	“S2”	“S3”
Final weight	0.17	0.50	0.33

In organizations with no financial constraints or low financial constraints: in military organizations or large organizations, there are low financial constraints, while at the same time the software should have various technical capabilities. Moreover, it should be able to fulfill all users’ needs since all their needs are of importance.

As indicated in the results, in organizations which have no special financial constraints or those which select the best option regardless of the cost issue, the final selection was “S.2” software. It should be noted that in the filtering stage, “S.2” software gained the highest score among other options, that means it has had more consistency with organization’s needs. In the investigated organization, the right for making choice was given to the management and both options were proposed with their conditions as a suggestion. The senior management of the organization selected “S.2” software as the final option after necessary evaluations.

5. Summary and conclusion

Organizations comprise big markets. In organizational markets, the number of shopping is less but their buying is huger and more complicated. When obtaining required services, organizational buyers are confronted with several decisions. These decisions depend on the situation of buying. Software buying is one of the organizational buying which is more complicated due to its nature. Thus, each solution which can help to lessen its complexity gains attention. This study deals with this problem and its aim is to provide a solution for organizational buying. After presenting the software selection model, determined some indicators for buying software and in a case study based on MADM methods evaluated and selected the comprehensive university system for the organization. The study’s main findings are as follows:

1.
Currently, the best method for evaluating and selecting software based on indicators is using MADM.
2.
The most suitable MADM methods for these problems are AHP and ELECTRE methods.
3.
The best choice might be different in different organizations and directly depends on its risks, needs and the budget.
4.
The cost criterion is the most influential parameter in software selection and also the most important organizational constraint.

6. Suggestions for future research

Software implementation of the work was not carried out; as a DSS of software selection, a system can be implemented.

In this study, values of all indicators were identified by individuals. Some criteria can be obtained from spontaneous methods.

The number of case studies was somehow limited; the accuracy of indicators and models can be investigated in other organizations or other software.

Footnotes

Conflict of interest

None to report.

References

Ariyachandra,

Watson,

, (2010). Key organizational factors in data warehouse architecture selection, Journal of Decision Support Systems, 49, 200-212.

Bruce,

S.B.

, (1991). A Process Model of Make-vs.-Buy Decision-Making; The Case of Manufacturing Software, IEEE Transaction on Engineering Management, 38(1), 24-32.

Bernroider,

Obwegeser,

Stix,

, (2011). Analysis of heuristic validity, efficiency and applicability of the profile distance method for implementation in decision support systems, Journal of Computers & Operations Research, 38, 816-823.

Parvin,

Minaei-Bidgoli,

Alinejad-Rokny,

Punch,

W.F.

, (2013). Data weighing mechanisms for clustering ensembles, Computers & Electrical Engineering, 39(5), 1433-1450.

Burger,

P.C.

Cann,

C.W.

, (1995). Post-purchase strategy, Industrial Marketing Management, 24, 91-98.

Cortellessaa,

Marinellia,

Potena,

, (2008). An optimization framework for build-or-buy decisions in software architecture, Journal of Computers & Operations Research, 35, 3090-3106.

Figueira,

Greco,

Ehrgott,

, (2005). Multiple Criteria Decision Analysis: State of the Art Surveys, International Series in Operations Research & Management Science, 78, ISBN 978-0-387-23067-2.

Gupta,

, (2011). Material selection for thin-film solar cells using multiple attribute decision making approach, Journal of Materials and Design, 32(3), 1667-1671.

Guillermo,

Mendoza,

, (2000). Multiple criteria decision making approaches to assessing forest sustainability using criteria and indicators: a case study, Journal of Forest Ecology and Management, 131, 107-126.

10.

Parvin,

Minaei-Bidgoli,

Parvin,

, (2013). A new classifier ensemble methodology based on subspace learning, Journal of Experimental & Theoretical Artificial Intelligence, 25(2), 227-250.

11.

Gungor,

Baracli,

Selcuk,

Baslıgil,

, (2009). An integrated decision support system dealing with qualitative and quantitative objectives for enterprise software selection, Journal of Expert Systems with Applications, 36, 5272-5283.

12.

Ghingold,

Wilson,

D.T.

, (1998). Buying center research and business marketing practice: meeting the challenge of dynamic marketing, Journal of Business and Industrial Marketing, 13, 96-108.

13.

Galorath,

, (2008), Software Project Failure Costs Billions. Better Estimation & Planning Can Help, viewed 22-08-2017. http://www.galorath.com/wp/software-projectfailure-costs-billions-better-estimation-planning-can-help.php.

14.

Hwanga,

Kob,

Goanc,

, (2007). Web-based multi-attribute analysis model for make-or-buy decisions, Journal of Mathematical and Computer Modeling, 46, 1081-1090.

15.

Huang,

Zhang,

Liu,

Sutherland,

, (2011). Multi-criteria decision making and uncertainty analysis for materials selection in environmentally conscious design, Int. J. Adv. Manuf. Technol., 52, 421-432.

16.

Hwang,

C.L.

Yoon,

, (1981). Multiple Attribute Decision Making: Method and Applications, Springer Verbleg.

17.

Alavi,

Akbari,

Alinejad-Rokny,

, (2012). Plant type selection for Sarcheshmeh copper mine by fuzzy TOPSIS method, An International Jour. Advanced Engineering Technology and Application, 1(1), 8-11.

18.

Juha,

Pentti,

, (2008). Managing risks in organizational purchasing through adaptation of buying centre structure and the buying process, Journal of Purchasing & Supply Management, 14, 253-262.

19.

Kumar,

Agrawal,

V.P.

, (2009). Attribute based specification, comparison and selection of electroplating system using MADM approach, Journal of Expert Systems with Applications, 36, 10815-10827.

20.

Esmaeili,

Minaei-Bidgoli,

Alinejad-Rokny,

, (2012). A Hybrid Recommender System for Joining Virtual Communities, Research Journal of Applied Sciences Engineering and Technology, 4(5), 500-509.

21.

Moon,

Tikoo,

, (2002). Buying decision approaches of organizational buyers and users, Journal of Business Research, 55, 293-299.

22.

Mokhtari,

Jalalifar,

Alinejad-Rokny,

PourAfshary,

, (2011). Prediction of Permeability from Reservoir Main Properties using Neural Network, Scientific Research and Essays, 6(32), 6626-6635.

23.

Eftekhari,

Poyan,

Alinejad-Rokny,

, (2011), Evaluation and Classifying Software Architecture Styles Due to Quality Attributes, Australian Journal of Basic and Applied Sciences, 5(11), 1251-1256.

24.

Seyedaghaee,

Rahati,

Alinejad-Rokny,

Rouhi,

, (2013). An Optimized Model for the University Strategic Planning, International Journal of Basic Sciences & Applied Research, 2(5), 500-505.

25.

Seyedaghaee,

Amirgholipour,

Alinejad-Rokny,

Rouhi,

, (2012). Strategic Planning Formulation by Using Reinforcement Learning, Research Journal of Applied Sciences, Engineering and Technology, 4(11), 1448-1454.

26.

Platts,

K.W.

Probert,

D.R.

Cahn,

, (2002). Make vs. buy decisions: A process incorporating multi Attribute decision-making, Int. J. Production Economics, 77, 247-257.

27.

Rainer,

Hall,

Baddoo,

, (2003). Persuading Developers to Buy into Software Process Improvement: Local Opinion and Empirical Evidence, Journal of International Symposium on Empirical Software Engineering (ISESE’03).

28.

Robinson,

P.J.

Faris,

C.W.

Wind,

, (1967). Industrial Buying and Creative Marketing. Allyn & Bacon, Boston, MA.

29.

Mokhtari,

S.M.

Alinejad-Rokny,

Jalalifar,

, (2014). Selection of the Best Well Control System by Using Fuzzy Multiple-Attribute Decision-Making Methods, Journal of Applied Statistics, 41(5), 1105-1121.

30.

Saaty,

T.L.

, (2008). Decision Making for Leaders: The Analytic Hierarchy Process for Decisions in a Complex World, Pittsburgh, Pennsylvania: RWS Publications, ISBN 0-9620317-8-X.

31.

Triantaphyllou,

, (2000). Multi-Criteria Decision Making: A Comparative Study, Dordrecht, the Netherlands: Kluwer Academic Publishers, pp. 320.

32.

Venkateswara,

Baral,

, (2011). Attribute based specification, comparison and selection of feed stock for anaerobic digestion using MADM approach, Journal of Hazardous Materials, 186, 2009-2016.

33.

Verville,

Halingten,

, A six-stage model of the buying process for ERP software, Journal of Industrial Marketing Management, 32, 585-594.

34.

Water,

Peet,

, (2006). A decision support model based on the Analytic Hierarchy Process for the Make or Buy decision in manufacturing, Journal of Purchasing & Supply Management, 12, 258-271.

35.

Wybo,

Robert,

Leger,

P.M.

, (2009). Using search theory to determine an applications selection strategy, Journal of Information & Management, 46, 285-293.

36.

Webster,

F.E.

Wind,

, (1972). Organizational Buying Behavior, Prentice-Hall, Englewood Cliffs, NJ.

37.

Borhani-Fard,

Minaei,

Alinejad-Rokny,

, (2013). Applying Clustering Approach in Blog Recommendation, Journal of Emerging Technologies in Web Intelligence, 5(3), 296-301.

38.

Zeleny,

, (1981). Multiple Criteria Decision Making, McGraw-Hill, New York.

Using multi-attribute decision-making approaches in the selection of a hospital management system

Abstract

Keywords

1. Introduction

2. Literature review

2.1 Organizational buyer behavior

2.2 Organizational software selecting indicator

Table 1 Indicators’ valuation based on two-paired comparison

1 Analytic Hierarchy Process.

Table 2 Software selection indicators

4.2 Implementation stage

4.2.1 First stage (filtering)

Table 3 Filtering items based on sufficiently good method

Table 4 Two-paired comparison of take log, track events and improve matters (A1)

Footnotes

Conflict of interest

References

Table 1
Indicators’ valuation based on two-paired comparison

¹
Analytic Hierarchy Process.

Table 2
Software selection indicators

Table 3
Filtering items based on sufficiently good method

Table 4
Two-paired comparison of take log, track events and improve matters (A1)