Optimizing Occupational health hazards using meta heuristics

Abstract

The stressfulness of manual activities of an occupational worker is evaluated by comparing job-demands to human-abilities or norms established in the scientific literature. If demands of a job exceed capabilities or published norms, it is perceived to be a severe job as it leads to risk of occupational health hazard (RoOHH). Departures from the capabilities or the norms are not uncommon in a developing country like India wherein an occupational worker is available at lesser cost with limited job opportunities. Moreover, a worker, in general, aims at maximizing his earnings by subjecting himself to extreme work conditions due to economic reasons. Factors like long working hours, improper rest breaks etc. severely affect workers health. Therefore we develop comprehensive and novel systems which not only depict the effects of extreme conditions on the health of the workers but also optimally tradeoff the conflicting objectives viz. maximizing the earnings and minimizing RoOHH in labor intensive manufacturing (LIM) units. Jobs having a high RoOHH, referred to as high risk jobs (HRJs), are usually high earning jobs. We devise a job-combination approach (JCA), wherein a worker doing a HRJ also does a low risk job (LRJ) to reduce his/her exposure to RoOHH. We quantify RoOHH in terms of a score for a given job-combination. It has been termed as risk assessment score (RAS) in one situation and composite discomfort score (CDS) in another situation. RAS/CDS is a direct indicative of perceived discomfort level (PDL) of workers. In our initial survey we identify that workers in LIM units in general aim to maximize their earnings by exposing them to high RoOHH due to reasons already mentioned. We also identify three factors which influence RAS/CDS of a given job, viz. number of working hours (WH), time of a rest break in minutes (RB), and number of rest breaks (NRB). RAS-earning tradeoff (RET) / CDS-earning tradeoff (CET) belongs to a class of combinatorial multiobjective optimization (MOO) problems.