1.1. Background
It is becoming a trend for enterprises to team up with a view of enhancing their competitive edge over their peers in the marketplace and to adapt to the technological innovations that are rapidly changing the business landscape. In these marketplaces, the consortium of these enterprises can effectively handle new business opportunities. A virtual enterprise (VE) is an alliance of autonomous entities (individuals, businesses), that have a common goal and work together, in a collaborative way. To achieve the goal, they collaboratively employ their varied competences and resources. It is of critical importance that business to business (B2B) relationships are effective as they enable the organizations to enhance their own ability to be more competitive. This research focuses on supporting B2B collaborations by using software agents. A crucial competitive factor of a VE is its ability to form an alliance that is customer focused.
In slightly over a decade, there have been a number of VE implementation models being proposed as highlighted in (Camarinha-Matos, Afsarmanesh, & Ollus, 2005; Romero & Vernadat, 2016). Industrial applications of VEs have been developed for the following industries: automotive industry, electronics and telecommunications, IT, building construction, transportation, chemical, food, textile, moulds among others. Manufacturing and supply chain are typical problems solved by VEs.
The lifecycle of a VE is divided into three phases (Pego-Guerra, 2006): 1) formation, 2) management and 3) dissolution. The formation phase establishes the goal and the objectives of the VE, according to the product demand. It also identifies the functional requirements that organization needs to fulfil. After the functional requirements are known, the core capabilities needed by VE are determined. Several enterprises may have these core capabilities, but only few of them are selected as members of the VE. This process is defined as the partner evaluation and selection process. Once the partner evaluation and selection process is finished, the VE enters its management phase. The management phase focuses on how to achieve the goals and objectives of the VE. In the management phase, members collaborate and integrate their core competencies to satisfy the functional requirements, identified in the formation phase. The performance of partners is also evaluated in this phase. Finally, once the product demand is met, the VE dissolves, and its members find other value-adding chains, where their core capabilities can be used. The dissolution phase deals with ending the relationship among partners and eventually the evaluation of the results of the collaborative work.
One of the approaches applied in modelling VEs is the use of multi-agent systems approach (Musumba & Wamuyu, 2016). There are many past works focusing on agent-based approaches and systems. However, they do not address a VE model holistically. They simplify the concept of a VE and focus on a single aspect or a single phase in the lifecycle of a VE. Existing works also lack solutions that are flexible and applicable across diverse sectors. Existing models lack dynamism and are incapable of handling varied VE requirements in diverse domains. Various partner selections and collaboration processes exist but are non-adaptive to diverse domains. They are also human resource intensive and costly. Integration of human negotiation capabilities in the VEs, which are key in forming practical consortia is seldom.
This work proposes a systematic approach to modelling VEs so that various forms of VEs can be realized thus allowing flexibility. This is demonstrated by the use of case studies from varied domains. This study spans software agents’ technology and virtual enterprises. Software agents’ technology is considered to be a special branch of distributed artificial intelligence (DAI). This study uses case scenarios to demonstrate the applicability and possible evaluation of the model.