The problem `human and work in a model working group is investigated by means of cellular automata technique. Attitude of members of a group towards work is measured by an indicator of loyalty to the group (the number of agents who carry out their ta
sks), and lack of loyalty (the number of agents, who give their tasks to other agents). Initially, all agents realize scheduled tasks one-by-one. Agents with the number of scheduled tasks larger than a given threshold change their strategy to unloyal one and start avoiding completing tasks by passing them to their colleagues. Optionally, in some conditions, we allow agents to return to loyal state; hence the rule is hysteretic. Results are presented on an influence of i) the density of tasks, ii) the threshold number of tasks assigned to the agents forcing him/her for strategy change on the system efficiency. We show that a `black scenario of the system stacking in a `jammed phase (with all agents preferring unloyal strategy and having plenty tasks scheduled for realization) may be avoided when return to loyalty is allowed and either i) the number of agents chosen for task realization, or ii) the number of assigned tasks, or iii) the threshold value of assigned tasks, which force the agent to conversion from loyal strategy to unloyal one, or iv) the threshold value of tasks assigned to unloyal agent, which force him/her to task redistribution among his/her neighbors, are smartly chosen.
Spread of information in crowd is analysed in terms of directed percolation in two-dimensional spatial network. We investigate the case when the information transmitted can be incomplete or damaged. The results indicate that for small or moderate pro
bability of errors, it is only the critical connectivity that varies with this probability, but the shape of the transmission velocity curve remains unchanged in a wide range of the probability. The shape of the boundary between those already informed and those yet uninformed becomes complex when the connectivity of agents is small.
