A Dynamic Job Shop Model for Scheduling Tasks in a Software Development Environment

Abstract

This work proposes a Dynamic Job Shop Scheduling (DJSS) model for scheduling the task in a software
production environment. The aim is to organize the insertion of tasks in the shop in a precedence order
previously defined that obeys the mapping of the critical path and the critical chain to identify constraints
in the productive process. The complexity and dynamism of the software development environment
require a mature and adjusted process that enables management throughout its extension. Therefore, the
Unified Process (UP) was used in conjunction with the Theory of Constraints (TOC) in the Software
Development Process (SDP). The UP requires an efficient model for the processing of activities, given
by the set of independent variables involved in the process. For this reason, a model to programming
the task based on dynamic scheduling was developed. The problem around the DJSS is to program
the tasks in the shop in a way that allows identifying one or more production lines in the shop with
constraints of capacity at runtime. A capacity constraint is any element that disturbs the productive
process, causing L={Mi Aj} to have delivery (DA) of less than 100% of the artifacts, given by the set
of jobs (jn
) scheduled. The resource with capacity constraint limits the production of the machines in the
production line, leading to bottlenecks in the production process. A bottleneck, in turn, is denoted by
the production limitation of a machine (people) caused by one or more capacity constraints that imply
maximum production, which in this case is defined by D={Mi, Aj=100%}. In this work, the maximum
production is given by the set of tasks processed in each of the phases of the UP and according to the
productive calendar, which delimits the deadline of each delivery and establishes the budget of hours
for the shop. The set of tasks scheduled and “rescheduling” and properly processed in all phases define
the total production of each machine. The use of DJSS facilitated the organization and management of
insertion and rescheduling operations in the shop and as result, produced data to measure the effort/time
that allowed identifying the bottleneck of the software production process.