Software for problem solving
Complex problems may require simultaneous consideration of many performance criteria. Two new fields, multi-objective optimisation and evolutionary multi-objective optimisation (EMO), can readily contribute to such problem solving. The EU-funded NH-MCDM (New horizons for multi criteria decision making) project worked on applying the above optimisations to a new field of MCDM. It aimed to develop an MCDM framework for solving complex problems based on the use of optimisation software. In addition, the team worked on developing a flexible and structured approach for managing the design process. This was achieved through a consortium comprising one EU partner plus members from Brazil, Canada, Israel and Mexico. One of the project's key objectives was dissemination. In this context, the consortium organised information exchange through several conferences, courses and workshops, plus seminars and tutorials, during which the team examined current studies in EMO and considered new directions. Moreover, the project advanced EMO theory through research into a set of novel applicable algorithms, several of which boasted specific advantages in furthering EMO. One important outcome was the development of a new methodology for optimising products for both single- and multiple-objective problems, enabling products to adapt to changing conditions. The team then researched applications concerning multi-objective games. Outcomes suggest evolutionary algorithms for such games and illustrate a new equilibrium that supports players’ strategizing and decision-making. NH-MCDM also explored ways of enhancing robotic responses to uncertain scenarios through mechanical cognitivisation. The concept essentially means robot training, and initial proof-of-concept shows potential for this research direction. Overall the project led to a new understanding of MCDM as a means to solving complex problems using novel software. The work has also led to new applications in robot cognition.
Keywords
MCDM, EMO, multi-objective games, mechanical cognitivisation, robot cognition
