A project portfolio selection framework for transforming Iranian universities into entrepreneurial institutions

Purpose – This research proposes a framework by which universities can define and implement projects that transform them into entrepreneurial universities. The frameworkhelps decision-makers identify suitable goals and strategies, gather a list of projects to fulfill the goals and strategies and prioritize the projects and form a portfolio. Design/methodology/approach – In the proposed framework, importance–performance matrix, hierarchical strategic planning, Delphi technique, DEMATEL-based ANP and a multi-objective model are used. The mathematical model consists of four objective functions including efficiency, quality and balance maximization and also cost and risk minimization. The proposed framework is applied to Amirkabir University of Technology, Tehran, Iran, and the results are brought in this paper. Findings –The output of the proposed framework is a portfolio of projects that aims to transform a traditional university into a third-generation one. Although the final portfolio must be customized for different universities, the proposed steps of the framework can be helpful for almost all cases. Originality/value – The suggested framework is unique and uses both qualitative and quantitative techniques for project portfolio selection.


Introduction
Iran's economy has relatively low productivity. The annual budget of the country is mostly dependent on the sale of raw materials. Also, transfer of capital and technology from the developed countries to Iran is associated with many difficulties (World Bank, 2017). There is a large number of university graduates looking for a job in Iran, and the market is incapable of creating enough new job opportunities for them. Also, the education system does not teach entrepreneurial and self-employment skills to the students (Almonitor, 2015). Most of the economists believe that the technology-based entrepreneurship and adding knowledge and creativity to businesses and products are the best solutions for improving the economic situation of Iran (Karimi et al., 2010). In the near future, those countries which have proper infrastructure for creating knowledge-based added value will be the main players in the global economy (Vrateovska et al., 2014). Due to the increasing demand for self-employment training, the need for establishing entrepreneurial universities is more than ever before (Alexander and Evgeniy, 2012). The most important and influential producers of knowledge are universities. Commercializing the knowledge of universities has several benefits for the economy. It increases the rate of job creation, improves the competitiveness of the country in the world, stops the brain drain phenomenon in the developing countries' domestic production, prevents the sale of natural resources in raw form and decreases the outflow of capital for importing technologies (Kalar and Antoncic, 2015). An entrepreneurial University is the place for training expert human resource. These experts should be systematically able to transfer their ideas and thoughts to the society. The society should also be able to use these new ideas. Many of the country's issues and challenges, including security, employment, healthcare, cultural development and social welfare require new and technological ideas urgently. Also, the authorities need scientific and professional solutions for social, economic and environmental problems of the country. University is the best source for obtaining these solutions (Sooreh et al., 2011).
But, developing an entrepreneurial university is a very complex and time-consuming program. In such a situation, failure will be inevitable without a strategic plan and a portfolio selection framework. This study aims to develop a comprehensive framework to help traditional universities move toward third-generation universities. The proposed framework answers two important questions including the following: 1. How to set the strategic goals and priorities? 2. How to define a portfolio of projects and programs to reach strategic goal?
2. Literature review and theoretical framework Entrepreneurial universities are one of the most important constituents of knowledge-based economies. After the second academic revolution, the third mission was added to teaching and research which were the traditional missions of the university. This new mission is entrepreneurship and the universities that have developed the needed infrastructures for this mission are called entrepreneurial universities. Etzkowitz described the third-generation university as an institution that has many research contracts and strategic partnerships with other organizations and is financially independent (Etzkowitz, 1984). Clark insisted on the concept of innovation as a key feature of the third-generation universities (Clark, 1998). Chrisman added the idea of founding spin-offs by academics, graduates and students to the main concept of the entrepreneurial university (Chrisman, 1995). Dill focused on the commercialization of university researches and suggested university technology transfer (UTT) (Dill, 1995). According to Ropke, an entrepreneurial university should have three features: It should be an entrepreneurial institution, provide its members with knowledge and skills needed for being an entrepreneur and set its relationships with other entities based on the policy of entrepreneurship (R€ opke, 1998). Sporn paid attention to the interaction of entrepreneur universities with their environment and determines how they should adapt to the dynamic environmental situation (Sporn, 2001). Poole listed the failure and success factors of the international strategy of an entrepreneurial university (Poole, 2001). Audretsch et al. conducted a comprehensive research about entrepreneurial finance and technology transfer. The study is focused on the role of governmental venture funds, subsidy programs and patent-friendly regulatory. They also discussed the impact of technology transfer offices on university entrepreneurship and regional competitiveness (Audretsch et al., 2016). Zhoa et al. introduced four characteristics that can be seen in all of the entrepreneurial universities (Zhoa, 2004): (1) Revenue generation through the transfer of knowledge and technology and the sale of patents.
(2) Considerable influence on regional industries and economy.
(3) Adoption of the entrepreneurship ideology among the academics. motivations that can stimulate academics to follow the mission of entrepreneurship. They concluded that the managers should not exclusively focus on the monetary incentives and it is better to consider a wider range of incentives to improve the level of interaction between academia and industry (D'Este and Perkmann, 2011). Jacob et al. claimed that the success of the transformation toward entrepreneurship depends on the national climate and the internal policies of the universities. They mentioned the infrastructural and cultural changes needed for this gradual transformation process. They also pointed out that the universities face a kind of role uncertainty when they start to carry out the third mission. Flexibility and diversity in both macro and micro levels are necessary for solving this issue (Jacob et al., 2003). Kirby listed seven critical barriers to entrepreneurship in universities as follows: (Kirby, 2006) (1) Relationships are typically impersonal.
(2) Structure of the universities is hierarchical and bureaucratic.
(3) Rules and procedures are constraining and anti-creative.
(5) The need for immediate results contradicts the time-consuming nature of becoming entrepreneurial.
(7) Lack of a suitable strategic plan and road map.
He also proposed some strategies to overcome the mentioned barriers. These strategies are summarized in Table I: Guerrero et al. categorized the factors affecting the development of entrepreneurial universities into formal and informal factors. They also proposed a set of indicators and measures to ease monitoring and assessing the transformation process toward entrepreneurship. Table II shows the overall form of the assessment method (Guerreo et al., 2006). According to Nelles & Vorley, five critical elements form the fundamentals of an entrepreneurial university. These elements are shown in Figure 1 Salamzadeh et al. claimed that for transforming to entrepreneurship, universities should revise and improve their processes. The most important processes which must be considered are teaching, research, managerial, logistical, commercialization, selection (for students, university professors and staff), funding and financial, networking and multilateral interaction processes (between students, university professors, staff, industrial Unique framework for Iranian universities Action Activity

Endorsement
Top managers and high-ranking staff should act as role models Incorporation Faculty, department and personal plans Implementation Setting targets and monitoring them Communication Consulting on the strategies and disseminating them Encouragement and support Hard support: laboratories, pre-incubators, incubators, science parks, meeting rooms, computing support, office support services and seed corn funding Soft support: training, mentoring and advice, signposting to sources of external support and ongoing technical and management support once the venture is launched Recognition and reward Fairness, job promotion, etc. Organization Cross-disciplinary research and teaching groups, educational partnerships, a multidisciplinary Entrepreneurship Centre Promotion Business plan competitions, entrepreneurship "halls of fame", Cases, role models Table I. Strategic actions to overcome the barriers to entrepreneurship researchers, entrepreneurial centres, industries, policy makers and society) . Rhoades expanded Clark's theories about entrepreneurial university by relating the considerations of systems analysis and organizational studies (Rhoades, 2017). According to Etzkowitz, becoming an entrepreneurial university takes place in three stages: University entrepreneur one: The university must determine its strategic direction, start acquiring the needed abilities, develop a facilitative legal framework and set its own priorities.
University entrepreneur two: The research activities of the members must actively get commercialized. Facilitating the technology transfer, enhancing the research corporations, preserving the intellectual properties and supporting the start-ups are also the important tasks of this stage.
University entrepreneur three: The university takes a leading role in innovation, has a tight relationship with the regional industries and government and makes a significant contribution to both regional and national economy (Etzkowitz, 2016). In this research, initial list of projects is created based on the Etzkowitz model.

Methodology
This study aims to propose a project portfolio selection framework that facilitates the transition toward university entrepreneurship. This framework consists of five stages including the following: Stage 3: Identifying the relationships between the final-level goals and rankings the goals.
Stage 4: Creating a list of candidate projects and programs to meet the objectives.
Stage 5: Creating a portfolio of projects by a multi-objective mathematical model.

Evaluating the status of the academic entrepreneurship indicators
The two main questions raised at this point are as follows: What indicators should be evaluated? and how should this evaluation be done?
In this research, we evaluate the indicators of Table II which are suggested by Guerrero et al. (Guerreo et al., 2006), by the importance-performance analysis (IPA) which was firstly introduced by Martilla & James (Martilla and James, 1977). IPA is a gap analysis method. Data collection in this technique is very similar to the SERVQUAL technique. IPA is an effective tool for evaluating the competitive position of an organization, identifying development opportunities, designing marketing strategies and providing targeted services. For the first time, this method was used to identify and prioritize the product or service characteristics that the organization should focus on to maximize its customer satisfaction. Through the formation of a two-dimensional matrix its vertical axis is performance (quality) of each feature and its horizontal axis is the importance of that feature in customers' decisionmaking. Then, a two-dimensional network is created which consists of four areas. Figure 2 shows the importance-performance matrix and corresponded strategies. The measurement method is as follows: The questionnaires are designed according to the indicators of Table II to evaluate the performance of the university based on each indicator. Validity and reliability of the questionnaires are checked by the experts and Alpha Cronbach test respectively. Also, the importance of the indicators can be determined by one of the multi-criteria decision-making (MCDM) methods. After quantifying the performance and importance of the indicators, critical levels must be determined for performance and importance. If the importance degree of an indicator is larger than the critical level, it should be considered a high importance indicator. Otherwise, it will be considered a low importance indicator. The same process should be carried out for the performance dimension. In the next stage, we focus on the indicators which are located in quarter four of the importance-performance matrix. Of course, this does not mean that other indicators should be ignored. These indicators should also be developed based on their own strategies in the importance-performance matrix.

Determining the strategies and goals
Indicators that are located in the fourth quarter are the best choices for improvement and investment. The universities should set strategies and goals to improve the selected indicators effectively. In this study, the hierarchical strategic programming with zigzag motions is used (Shirazi, 2005). In the literature of strategic management, the goals and strategies are usually presented in a hierarchical structure. However, these structures are usually not completely relevant to one another and their dependency is not clear. Although major studies classify the objectives into long-term, mid-term and short-term, in this study, organizational goals are divided into two main levels as follows: 1. Major goals 2. Operational goals.
Major goals: This includes the goals that the university wants to achieve in the long term. In other words, there is idealism in expressing these goals. For example, the goal is to increase profits, quality, credibility and so on, which are important for the university. In order to JIUC Figure 2. The importanceperformance matrix and the corresponded strategies achieve the major goals, the university must determine the major strategies eg. the development strategy, diversity strategy and so on.
Operational goals: These goals are usually expressed at different levels of long term, mid term and short term. These goals reflect the expected results of the strategies of the previous level. Operational goals should be quantitative, measurable, realistic, understandable, challenging, hierarchical, achievable, and consistent with other organizational goals. The hierarchical strategic planning with zigzag motion is based on the axiomatic design method and aims to create harmony between the two spaces of strategies and goals at all levels. In the hierarchical strategic planning, the mission and vision of the organization must be determined at first. Then, major organizational goals must be outlined. Then, at least one major strategy must be set for each major goal. At the next level, for the successful implementation of each strategy, at least one goal must be set. These goals should be in line with the higher-level goals. In fact, we move between the goals space and the strategies space with zigzag motions until the goals are specific enough to be met with one or two projects or activities. Figure 3 shows the overall form of the hierarchical strategic planning with zigzag motions.

Identifying the relationships between the goals and rankings them
The suggested projects and programs for creating and developing an entrepreneurial university are in line with the final-level goals. Therefore, the importance of these projects and programs is a function of the importance of their corresponded goals. Therefore, before listing the proposed projects, the existing relationships between the goals must be detected and the importance of each goal must be determined. In this study, the DEMATEL-based ANP (DANP) method is used to do this.
The DANP method is one of the multi-criteria decision-making (MCDM) methods. It computes the ANP super matrix using the DEMATEL communication matrix and calculates the weights of criteria and sub-criteria. In fact, the DANP method is the hybrid form of DEMATEL and ANP. This method has nine steps (Chiu et al., 2013): Step 1: calculate the direct influence matrix by scores The relationships between the goals (influence of a goal on other goals) are expressed based on the experts' opinions using a five-point scale (0-4). 0 5 no influence, 1 5 low influence, 2 5 medium influence, 3 5 high influence and 4 5 very high influence . Thus, the direct influence matrix (D) can be calculated. (Eqn 1) Unique framework for Iranian universities (1) Step 2: normalizing the direct influence matrix The normalized direct influence matrix can be obtained using Eqn 2: Step 3: calculating the total influential matrix (TC) The total influential matrix is obtained by Eqns 3 and 4. Note that "I" represents the unit matrix.
Step 4: analyze the results In this step, the summations of each row and column should be calculated separately according to Eqns 5 to 6.
T ¼ t ij ½ ; i; j ∈ f1; 2; . . . ; ng The index ðr i Þ represents the sum of the rows ðiÞ and the index ðc j Þ represents the sum of the column ðjÞ. The index ðr i þ c j Þ is obtained from the sum of the row ðiÞ and the column ðiÞ and shows the importance of the criteria ðiÞ. Similarly, the index ðr i − c j Þ shows how much the criterionðiÞ affects the other criteria and gets influenced by them. If ðr i − c j Þ is positive, the JIUC criterion ðiÞ affects some of the other criteria, otherwise it gets influenced by some of the other criteria. Now we can use DANP for finding the influential weights in each criterion: Step 5: find the normalized total influential matrix The normalized form of the matrix TD is obtained from the mean T c ij ½ . Thus, the sum of each row is computed, and each element is divided by the sum of the elements of its corresponded row. (Eqns 8 and 9) Step 6: find the normalized form of ðT C Þ by dimensions and clusters In this step, matrix T C is normalized with the total degrees of effect and influence of the dimensions and clusters.
Unique framework for Iranian universities Step 7: building an unweighted supermatrix W C The transposed form of the matrix T α c is called "unweighted supermatrix" and is shown by W, as in Eqn 12.
Step 8: building a weighted supermatrix ðW α Þ The weighted supermatrix is obtained by the product of the normalized total influential matrix ðT α D Þ and the unweighted supermatrix (W). (Eqn 13) Step 9: find the influential weights of the DANP The weighted supermatrix must be raised to a sufficiently large power Z until it converges and reaches stability. The output of this step is the effective DANP weights. (Eqn 14) 3.4 Creating a list of candidate projects and programs to meet the objectives The most important issue in defining a project is scope management. The scopes of the proposed projects should be consistent with at least one of the last-level goals set out in Section 3.2. The high-level scope of the project must be outlined in the project charter (according to BMBOK) or the business case (according to Prince2) (PMBOK, 2017) (Prince2, 2017). By specifying the scope of the project, we can see what the project contains or does not contain. It also shows with what goals the project is consistent, on which goals have negative effects and on which ones has no impact. In this study, the Delphi technique is used to identify and sort the most important projects and programs for entrepreneurship. Although the Delphi technique is not a MCDM method, it can be used before applying these techniques to reach an agreement on the candidate projects. Figure 4 denotes the main steps of this method and their sequence.
In the first stage, the problem statement is identified, and the characteristics of the experts' panel members are determined based on that. Then, the qualified experts are identified and invited to the panel. Regarding the subject of this research, top managers, high-level university staff, faculty members, entrepreneurship experts, managers of the key industries, government officials, spin-off owners and entrepreneur students can be members of the experts' panel.
The second step of the Delphi method is to generate ideas in the field of research. Expert panel members express their views about the questions the researchers ask. By analyzing and refining these ideas and removing duplicates and identical terms, the researcher extracts the list of issues related to the topic of research. The proposed method of this study is to provide the expert panel members with the last-level goals outlined in Section 3.2 and ask them to suggest programs and projects achieve the goals. By editing the suggestions, the initial list of projects will be formed. In the third step, members are asked to express the importance of the items listed in the initial list by linguistic or numerical variables or select some of the most important ones. Then, using MCDM methods such as AHP, or the Q-Sort  Figure 4. The Delphi technique method, we exclude the projects or programs that are considered unimportant by the members. This process continues until the members reach a consensus on the list of suggested projects. The output of section 3.4 is a list of projects and programs suggested by the experts, in which the importance of each project is also specified.

Creating a portfolio of projects and programs by a multi-objective mathematical model
The projects with the highest degree of importance are not necessarily the best choices to carry out. In addition to maximizing benefits, the organization should consider other important factors such as risk, balance and budget constraints. In the following points, we propose a multi-objective programming model that can provide an optimal portfolio of projects and programs for moving toward entrepreneurship considering the mentioned factors. Compared to the similar models, this one is simpler and more comprehensive. As the model is simple and its parameters are definite, it can be solved with the typical software and by people who are not familiar with the operation research. Parameters: (1) f m ¼ Weight of the strategy S m (1) Z 1 chooses the projects whose total weights are maximal. Z 2 divides the budget between the projects in a way that the university can grow in a balanced way. To do this, the maximum productivity of each of the sub-portfolio or program is measured, and the objective function maximizes the least relative productivity among the sub-portfolios and programs. Z 3 maximizes the total quality level of the portfolio. Z 4 lessens the total cost of the portfolio. Z 5 minimizes the total risk score of the portfolio. For more simplicity, we can change Z 1 , Z 2 and Z 3 into restrictions. So, the model can be rewritten as follows: Max Unique framework for Iranian universities 4. Case study (Amirkabir University of Technology, Tehran, Iran) Amirkabir University of Technology (Tehran Polytechnic) is the first and most experienced school of engineering in Iran. Admission to this university is done through the national entrance exam and is considered very competitive. The university has developed nine strategies in three dimensions as follows: the content, structure and environment for becoming a third-generation university (Table III).
As mentioned earlier, these goals are not independent and can have positive or negative effects on each other. The direct influence matrix for these nine goals is shown in Eqn 36. (Based on the expert judgment).
After completing the DANP steps, the weight is obtained for each of the goals.   Also, to implement each of these nine strategies, there are a number of proposed solutions that can be seen in Table V. By collecting information about the project costs and the associated risks of them as well as the total budget of the organization and its maximum tolerable risk, it is possible to prioritize the projects and form the project portfolio. The costs and risks of projects can be expressed in deterministic, fuzzy, grey numbers or probabilistic form. Depending on the conditions and type of projects, other constraints can be added to the proposed model of this research. If a part of the projects can be done and another part can be postponed to subsequent periods, continuous variables can be used instead of binary variables.

Conclusion
Because the entrepreneurial university (third generation) is one of the core requirements of the knowledge-based development, all universities need to have a comprehensive plan to transform them into entrepreneurial universities at the lowest possible cost and risk. The need for such a plan is more vital in developing countries, such as Iran, due to poor industry and poor university connectivity. This paper suggests a framework that help universities make such a plan. This framework consists of five steps including evaluating the status of the academic entrepreneurship indicators, determining the strategies and goals, identifying the relationships between the goals and rankings them, creating a list of candidate projects and programs to meet the objectives and creating a portfolio of projects and programs by a multi-objective mathematical model. The output of this framework is a portfolio of projects that according to the budget and other university conditions have the highest priority in terms of transforming traditional universities into entrepreneurial universities.