Approaches on Generating Optimized Sequences of Items Used in Assessment.

Assessment is a key element in education. Given its multiple methods, assessment contains the aspect of correctness and honesty from both the assessor and the assessed. These methods can be classified and their types are numerous, starting from single questions to tests and practical work. Concerning tests, their vast usage in practice raises problems such as progressing difficulty, concept understanding regardless the test morphology or test sort depending on subject. Generally, we will refer at questions within a test or tests within a group of tests as items. In this paper, we present several approaches for generating items, defining or not some restrictions, in order to obtain optimized sequences of items (OSI). These restrictions consist in connections between items regarding degrees of difficulty or sorting depending on item subject. These cases will be solved using nature-inspired algorithms (genetic algorithms), arborescent structures and lexical usage (keywords).