**1. Introduction**

Worldwide, the impacts of various anthropogenic activities, such as increased industrialisation, pollution, deforestation, and overconsumption, are causing destruction and overexploitation of natural resources and have been recognised as the most significant risk not only for the environment but also

for human health and well-being [1–3]. The overexploitation of natural resources has grown rapidly in the last two decades, and global resource supply chains have become extremely complex. This has resulted in increasing environmental pressures and impacts [2]. If humankind continues to live on the edge of or outside of the ecological limits, it will be much more di fficult to achieve equity, justice, prosperity, well-being, and healthy quality of life for everyone at the global and local levels [4–6].

The need for humanity to remain within the safe operating space of planetary boundaries and the need to eradicate poverty and accelerate sustainable socio-economic development are linked by the concept of "safe and just space for humanity" [7–9]. Therefore, a global shift from a linear economy towards a circular economy is needed, in which socio-institutional change along with resource efficiency and innovative product design [10,11] contribute to economic development and human well-being, with reduced pressures and impacts on the environment. The circular economy approach aims at continuous economic development to achieve waste minimisation, energy e fficiency, and environmental conservation, without posing significant challenges to the environment and natural resources [12]. The circular economy is a competitive environmental strategy to production processes and economic activity that allows resources to maintain their highest value while benefiting the business and society as a whole with better supply chain, low volatility of resource prices, better customer relations, improved services, and new employment opportunities [12,13]. The key considerations in the implementation of a circular economy are to refuse, reduce, rethink, repair, restore, remanufacture, and reuse resources [10,11,14], and pursue longevity, renewability, replaceability, and upgradability for resources and products that are used.

The sustainable development goal "one" of the United Nations seeks to eradicate poverty and subsequent inequalities in all forms, leaving no one behind. Owing to the multi-dimensional nature of poverty, its eradication involves complex interactions within socio-economic systems. Across the world, several development projects [15,16] and poverty alleviation programmes [17,18] have been implemented, which are primarily aimed at socio-economic development and poverty reduction of poor and vulnerable communities. The *Deendayal Antyodaya Yojana*–National Rural Livelihoods Mission (DAY-NRLM), a centrally sponsored programme of the Government of India, is one of the world's largest poverty eradication programmes that aims to eliminate rural poverty in the country through the promotion of multiple livelihoods for each rural poor household. The programme follows a participatory and community-demand-driven approach that focusses largely on the eradication of poverty and pays special attention to the development of social resilience and sustainable socio-economic development.

This study tried to identify an ideal strategy for sustainable socio-economic development planning for rural communities in developing economies by evaluating the DAY-NRLM programme. For this, we used fuzzy cognitive mapping—a participatory modelling approach that allows for the exploration and managemen<sup>t</sup> of various interrelationships between human, environmental, and socio-economic systems. The main principle behind using participatory modelling is that the resulting policy can effectively support a range of social, economic, and environmental goals. The systems-based integrated and participatory modelling approaches allow the stakeholders to contribute to the development of a system model, as well as support decision-makers' understanding of the system being examined and all its aspects, by participating in the scenario development process and identifying correct strategies within the domain. Therefore, it is safe to say that the contribution that stakeholders can o ffer to a given problem, usually related to support for decision-making, policies, regulation, or management, is valuable [19]. The fuzzy cognitive mapping approach allows stakeholders to collaboratively develop a "cognitive map" (i.e., a weighted, directed graph), which represents the perceived causal structure of their system [20]. Fuzzy cognitive map (FCM) [21] is a collection of concepts and causal relationships between them, in the form of a directed graph that can model a real-world system. Because FCMs can easily incorporate human knowledge and adapt to any given domain [22], they have been extensively applied in many research fields and areas of application [23,24]. Their simple model structure gives them considerable awareness and broad research interest [25].

This study also explored a new method for aggregating individual FCMs using the application of ordered weighted averaging (OWA) operators. The aggregation of individual FCM models must be taken into account when a variety of sources need to be included in the modelling procedure of a given system [23]. Individual FCMs, constructed by experts and/or stakeholders, can be aggregated to produce a combined FCM that will incorporate the knowledge from all the di fferent experts and/or stakeholders involved in the FCM construction process. The main objective behind aggregating individual FCMs is to improve the reliability of the overall model and make them less susceptible to potentially inaccurate knowledge of a single expert and/or stakeholders or knowledge inconsistency among the participants. Regarding the combination of multiple FCMs into a single collective model, among the techniques that can be found in the related literature, two methods are widely used in real-life problems [26]—the weighted average and the OWA method introduced by Yager [27]. The weighted average method is considered as the benchmark method for FCM aggregation purposes, where the final FCM model is built by averaging numerical values for a given interconnection [28]. However, there are certain limitations to implementing this methodology. For example, when a large number of stakeholders/participants are asked to assign values on the relationships of a given system, significant deviations can arise between these values, which shows an inconsistency of knowledge among the participants that leads to an inaccurate overall FCM. On the other hand, the application of OWA operators in the aggregation of individual FCMs has a limited presence in the literature. In particular, OWA operators in an FCM framework were introduced by Zhenbang and Lihua [29], who highlighted the ability of OWA operators to simulate the various AND/OR relationships between the concepts and studied the OWA aggregation under di fferent conditions. An OWA operator based on distance was examined by Leyva-Vazquez et al. [30] to rank the scenarios depending on the decision-makers' risk preferences.

The study used the context of the DAY-NRLM programme, which predominantly focuses on livelihood enhancement through building self-managed and sustainable community institutions of "the rural poor women". These community institutions are expected to overcome their social, financial, and economic exclusion. The broader objectives of the programme involve social mobilisation, institution building, enrolment of women in social security schemes and entitlements, socio-economic inclusion, sustainable livelihoods, capacity building, promotion of economic stability, improvement of social resilience, and skill development, aimed at elimination of rural poverty. Ultimately these outcomes will lead to the reduced socio-economic poverty and improved quality of life.

The paper is structured as follows: Section 2 describes the problem along with the methods used in this study. This section includes in particular the basic principles of FCMs, the specifications of the OWA operators, and the proposed algorithm based on the weights of the OWA operator to aggregate FCM weights. This section also deals with the quasi-qualitative method for constructing the collective FCM from groups, whereas in Section 3, the aggregation results produced from the application of the OWA method in the developed FCMs, the sensitivity analysis, and the scenario analysis are provided. In the end, Section 4 summarises the conclusions of this study.
